Ecology

1.Hutchinson, G. E. Concluding remarks. Cold Spring Harb. Symp. Quant. Biol. 22, 415–427 (1957).
Google Scholar 
2.Wethey, D. S. Biogeography, competition, and microclimate: The barnacle Chthamalus fragilis in New England. Integr. Comp. Biol. 42, 872–880 (2002).PubMed 

Google Scholar 
3.Heikkinen, R. K., Luoto, M., Virkkala, R., Pearson, R. G. & Körber, J.-H. Biotic interactions improve prediction of boreal bird distributions at macro-scales. Glob. Ecol. Biogeogr. 16, 754–763 (2007).
Google Scholar 
4.Bøhn, T. & Amundsen, P.-A. The competitive edge of an invading specialist. Ecology 82, 2150–2163 (2001).
Google Scholar 
5.Barger, C. P. & Kitaysky, A. S. Isotopic segregation between sympatric seabird species increases with nutritional stress. Biol. Lett. 8, 442–445 (2012).PubMed 

Google Scholar 
6.Gosselink, T. E., Deelen, T. R. V., Warner, R. E. & Joselyn, M. G. Temporal habitat partitioning and spatial use of coyotes and red foxes in East-Central Illinois. J. Wildl. Manag. 67, 90 (2003).
Google Scholar 
7.Odden, M., Wegge, P. & Fredriksen, T. Do tigers displace leopards? If so why?. Ecol. Res. 25, 875–881 (2010).
Google Scholar 
8.Pickett, E. P. et al. Spatial niche partitioning may promote coexistence of Pygoscelis penguins as climate-induced sympatry occurs. Ecol. Evol. 8, 9764–9778 (2018).PubMed 
PubMed Central 

Google Scholar 
9.Navarro, J. et al. Ecological segregation in space, time and trophic niche of sympatric planktivorous petrels. PLoS ONE 8, e62897 (2013).CAS 
PubMed 
PubMed Central 
ADS 

Google Scholar 
10.Reif, J., Reifová, R., Skoracka, A. & Kuczyński, L. Competition-driven niche segregation on a landscape scale: Evidence for escaping from syntopy towards allotopy in two coexisting sibling passerine species. J. Anim. Ecol. 87, 774–789 (2018).PubMed 

Google Scholar 
11.Trego, C. T., Merriam, E. R. & Petty, J. T. Non-native trout limit native brook trout access to space and thermal refugia in a restored large-river system. Restor. Ecol. 27, 892–900 (2019).
Google Scholar 
12.Durant, S. M. Competition refuges and coexistence: An example from Serengeti carnivores. J. Anim. Ecol. 67, 370–386 (1998).
Google Scholar 
13.Parmesan, C. & Yohe, G. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42 (2003).CAS 
PubMed 
ADS 

Google Scholar 
14.Parmesan, C. Ecological and evolutionary responses to recent climate change. Annu. Rev. Ecol. Evol. Syst. 37, 637–669 (2006).
Google Scholar 
15.Freeman, B. G., Scholer, M. N., Ruiz-Gutierrez, V. & Fitzpatrick, J. W. Climate change causes upslope shifts and mountaintop extirpations in a tropical bird community. Proc. Natl. Acad. Sci. 115, 11982–11987 (2018).CAS 
PubMed 
PubMed Central 

Google Scholar 
16.Alexander, J. M., Diez, J. M. & Levine, J. M. Novel competitors shape species’ responses to climate change. Nature 525, 515–518 (2015).CAS 
PubMed 
ADS 

Google Scholar 
17.Elmhagen, B. et al. Homage to Hersteinsson and Macdonald: Climate warming and resource subsidies cause red fox range expansion and Arctic fox decline. Polar Res. 36, 3 (2017).
Google Scholar 
18.IPCC. Climate change 2014: Synthesis Report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. (2014).19.Spielhagen, R. F. et al. Enhanced modern heat transfer to the Arctic by warm Atlantic water. Science 331, 450–453 (2011).CAS 
PubMed 
ADS 

Google Scholar 
20.Fossheim, M. et al. Recent warming leads to a rapid borealization of fish communities in the Arctic. Nat. Clim. Change 5, 673–677 (2015).ADS 

Google Scholar 
21.Descamps, S. et al. Climate change impacts on wildlife in a High Arctic archipelago: Svalbard Norway. Glob. Change Biol. 23, 490–502 (2017).ADS 

Google Scholar 
22.Descamps, S., Strøm, H. & Steen, H. Decline of an arctic top predator: Synchrony in colony size fluctuations, risk of extinction and the subpolar gyre. Oecologia 173, 1271–1282 (2013).PubMed 
ADS 

Google Scholar 
23.Garðarsson, A., Guðmundsson, G. A. & Lilliendahl, K. Svartfugl í íslenskum fuglabjörgum 2006–2008. Bliki 33, 35–46 (2019).
Google Scholar 
24.Merkel, F. et al. Declining trends in the majority of Greenland’s thick-billed murre (Uria lomvia) colonies 1981–2011. Polar Biol. 37, 1061–1071 (2014).
Google Scholar 
25.Fauchald, P. et al. The status and trends of seabirds breeding in Norway and Svalbard. 84 (2015).26.Williams, A. J. Site preferences and interspecific competition among guillemots Uria aalge (L.) and Uria lomvia (L.) on Bear Island. Ornis Scand. 5, 113 (1974).
Google Scholar 
27.Guidelines for the treatment of animals in behavioural research and teaching. Anim. Behav. 83(1), 301–309. https://doi.org/10.1016/j.anbehav.2011.10.031 (2012).28.Luque, S. P. An Introduction to the diveMove Package. 56 (2007).29.Luque, S. P. & Fried, R. Recursive filtering for zero offset correction of diving depth time series with GNU R Package diveMove. PLoS ONE 6, e15850 (2011).CAS 
PubMed 
PubMed Central 
ADS 

Google Scholar 
30.QGIS Development Team. QGIS Geographic Information System. (Open Source Geospatial Foundation Project. http://qgis.osgeo.org, 2018).31.Fieberg, J. & Kochanny, C. O. Quantifying home-range overlap: The importance of the Utilization Distribution. J. Wildl. Manag. 69, 1346–1359 (2005).
Google Scholar 
32.Calenge, C. The package adehabitat for the R software: A tool for the analysis of space and habitat use by animals. Ecol. Model. 197, 516–519 (2006).
Google Scholar 
33.Geange, S. W., Pledger, S., Burns, K. C. & Shima, J. S. A unified analysis of niche overlap incorporating data of different types. Methods Ecol. Evol. 2, 175–184 (2011).
Google Scholar 
34.Lewis, S., Sherratt, T. N., Hamer, K. C. & Wanless, S. Evidence of intra-specific competition for food in a pelagic seabird. Nature 412, 816–819 (2001).CAS 
PubMed 
PubMed Central 
ADS 

Google Scholar 
35.Linnebjerg, J. F. et al. Sympatric breeding auks shift between dietary and spatial resource partitioning across the annual cycle. PLoS ONE 8, e72987 (2013).CAS 
PubMed 
PubMed Central 
ADS 

Google Scholar 
36.McFarlane Tranquilla, L. A. et al. Multiple-colony winter habitat use by murres Uria spp. in the Northwest Atlantic Ocean: Implications for marine risk assessment. Mar. Ecol. Prog. Ser. 472, 287–303 (2013).ADS 

Google Scholar 
37.Pratte, I., Robertson, G. & Mallory, M. Four sympatrically nesting auks show clear resource segregation in their foraging environment. Mar. Ecol. Prog. Ser. 572, 243–254 (2017).ADS 

Google Scholar 
38.Kokubun, N. et al. Foraging segregation of two congeneric diving seabird species breeding on St. George Island, Bering Sea. Biogeosciences 13, 2579–2591 (2016).ADS 

Google Scholar 
39.Barger, C. P., Young, R. C., Will, A., Ito, M. & Kitaysky, A. S. Resource partitioning between sympatric seabird species increases during chick-rearing. Ecosphere 7, e01447 (2016).
Google Scholar 
40.Huffeldt, N. P. & Merkel, F. R. Sex-specific, inverted rhythms of breeding-site attendance in an Arctic seabird. Biol. Lett. 12, 20160289 (2016).PubMed 
PubMed Central 

Google Scholar 
41.Kappes, M. A. et al. Reproductive constraints influence habitat accessibility, segregation, and preference of sympatric albatross species. Mov. Ecol. 3, 34 (2015).PubMed 
PubMed Central 

Google Scholar 
42.Benvenuti, S., Bonadonna, F., Dall’Antonia, L. & Gudmundsson, G. A. Foraging flights of breeding thick-billed murres (Uria lomvia) as revealed by bird-borne direction recorders. Auk 115, 57–66 (1998).
Google Scholar 
43.Hunt, G. L., Bakken, V. & Mehlum, F. Marine birds in the Marginal Ice Zone of the Barents Sea in late winter and spring. Arctic 49, 53–61 (1996).
Google Scholar 
44.Hein, C., Öhlund, G. & Englund, G. Future distribution of Arctic Char Salvelinus alpinus in Sweden under climate change: Effects of temperature, lake size and species interactions. Ambio 41(Suppl 3), 303–312 (2012).PubMed 
PubMed Central 

Google Scholar 
45.Mehlum, F., Watanuki, Y. & Takahashi, A. Diving behaviour and foraging habitats of Brünnich’s guillemots (Uria lomvia) breeding in the High-Arctic. J. Zool. 255, 413–423 (2001).
Google Scholar 
46.Frederiksen, M. et al. Seabird baseline studies in Baffin Bay, 2008–2013. Colony-based fieldwork at Kippaku and Apparsuit, NW Greenland. Report No. 110. (Aarhus University, DCE – Danish Centre for Environment and Energy, Roskilde, Denmark., 2014).47.Spagnolo, M. & Clark, C. D. A geomorphological overview of glacial landforms on the Icelandic continental shelf. J. Maps 5, 37–52 (2009).
Google Scholar 
48.Meier, W. N. et al. Arctic sea ice in transformation: A review of recent observed changes and impacts on biology and human activity. Rev. Geophys. 52, 185–217 (2014).ADS 

Google Scholar 
49.Gaston, A. J., Smith, P. A. & Provencher, J. F. Discontinuous change in ice cover in Hudson Bay in the 1990s and some consequences for marine birds and their prey. ICES J. Mar. Sci. 69, 1218–1225 (2012).
Google Scholar 
50.Grémillet, D. et al. Arctic warming: nonlinear impacts of sea-ice and glacier melt on seabird foraging. Glob. Change Biol. 21, 1116–1123 (2015).ADS 

Google Scholar 
51.Valdimarsson, H., Astthorsson, O. S. & Palsson, J. Hydrographic variability in Icelandic waters during recent decades and related changes in distribution of some fish species. ICES J. Mar. Sci. 69, 816–825 (2012).
Google Scholar  More

Mathematical model 1—the relationship among the bodyweight of the initial WSSV-infected shrimp, number of deaths, and death time distributionThe experimental data show the time course of death for the infected shrimp satisfies the Laplacian distribution (Supplementary Tables 2–4). The relationship of the bodyweight of the initial infected shrimp number of deaths and death time distribution could be expressed by a mathematical model and the establishment of the mathematical model as shown below.Suppose that one dead shrimp could infect (n) healthy shrimp at the same day. These (n) infected shrimp do not die simultaneously but on different days (time course). The value of (n) is related to the weight of the dead shrimps—larger dead shrimp can infect more healthy shrimps of the same body weight. Our experimental results (Supplementary Tables 2–4) show the death time course for these (n) infected shrimp satisfies the Laplacian distribution, as follows:$$begin{array}{c}pleft(tright)=left{begin{array}{c}{b{{exp}}}left(-frac{left|t-aright|}{{c}_{1}}right),tle a\ {b{{exp}}}left(-frac{left|t-aright|}{{c}_{2}}right),t > aend{array}right.end{array}$$
(1)
where (a) is the peak time of number of dead shrimps, (b) is the maximal death percentage, ({c}_{1}) is related to the mortality increases of the infected shrimps, ({c}_{2}) is related to the mortality decreases of the infected shrimp, (p(t)) is the percentage of infected shrimp that die at time (t). The open bracket “{“ in formula (1) means the function is represented by two parallel expressions as described previously.Based on the Supplementary Tables 2–4, we can determine the value of (a), (b), ({c}_{1}), and ({c}_{2}) by the least square estimation method. As different weight corresponds to different distribution of death time, we can compute the relationship of weight of death shrimps with corresponding (a), (b), ({c}_{1}), and ({c}_{2}) (Supplementary Table 25).We found the relationship of (w) with (a), or (b), or ({c}_{1}) or ({c}_{2}) is quadratic (Eq. 2), with the data in Supplementary Table 25, we have$$begin{array}{c}left{begin{array}{c} a= -0.0918{w}^{2}+0.8772w+3.3449\ b=0.0029{w}^{2}-0.0369w+0.5849;;\ {c}_{1}=-0.0186{w}^{2}+0.1739w+0.7063\ {c}_{2}=0.0002{w}^{2}+0.0108w+1.0827;;,end{array}right.end{array}$$
(2)
Using Model 1, we can predict the effects of different body weights of dead WSSV-infected shrimp through the ingestion pathway of WSSV-infected dead shrimp on the WSSV transmission rate.Mathematical model 2—the dynamic changes of healthy, infected, and dead shrimp during WSSV transmissionWe derived and established Model 2 to simulate the WSS transmission dynamics in cultured shrimp. Using Model 2, we predicted the dynamic changes of three states (healthy, infected, and dead shrimps) in cultured shrimp as influenced by the WSS epidemic with the following:Now we can develop a model for the spread and break out of WSS. For any given weight (w) of shrimps, let ({s}_{h}(t)), ({s}_{i}(t)), and ({s}_{d}(t)) be the number of healthy shrimp, infected shrimp and dead shrimp respectively at time (t). Let (I(t)), (d(t)) be the number of daily infected shrimp, daily dead shrimp, respectively, at time (t).According to infection process, the decrement of healthy shrimp is caused by their infection, therefore we have (frac{d{s}_{h}}{{dt}}=-I(t)). The quantity change of infected shrimp includes the infection of healthy shrimp and the death of infected shrimp, we have (frac{d{s}_{i}}{{dt}}=I(t)-d(t)). The increment of dead shrimp is caused by the death of the infected shrimp; thus we have (frac{d{s}_{d}}{{dt}}=d(t)). We obtain the following system of ordinary differential equations:$$left{begin{array}{c}frac{d{s}_{h}}{{dt}}=-Ileft(tright)hfill\ ,frac{d{s}_{i}}{{dt}}=Ileft(tright)-dleft(tright)\ frac{d{s}_{d}}{{dt}}=d(t)hfillend{array}right.$$
(3)
where ({s}_{h}left(0right)={s}_{{h}_{0}}), ({s}_{i}left(0right)={s}_{{i}_{0}}), ({s}_{d}left(0right)={s}_{{d}_{0}}) are as the initial value, at (t=0).In the above system of ordinary differential equations, quantity (I(t)) can be expressed as follows$$begin{array}{c}Ileft(tright)={min }left{n{s}_{d}left(tright),{s}_{h}left(tright)-alpha {s}_{{h}_{0}}right}end{array},$$
(4)
(d(t)) can be expressed as$$begin{array}{c}dleft(tright)={int }_{0}^{T}{min }left{n{s}_{d}left(t-tau right),{s}_{h}left(t-tau right)-alpha {s}_{{h}_{0}}right}pleft(tau right)dtau end{array}$$
(5)
where (n) is the number of healthy shrimp infected by one dead shrimp on the first day. (p(tau )) is the death percentage of the (n) infected shrimp on the (tau) days, (T) is the longest survival time of infected shrimp.Now we explain how to set up the formulas (I(t)) and (d(t)). In the expression of (I(t)), (n{s}_{d}(t)) is the number of daily infected shrimp at time (t). But as the number of healthy shrimp decreases, there may not be as many as (n{s}_{d}(t)) healthy shrimp to be infected. Therefore, (I(t)) is the minimum of (n{s}_{d}(t)) and ({s}_{h}left(tright)-alpha {s}_{{h}_{0}}), where (alpha (0 , < , alpha, < , 1)) represents the percentage of healthy shrimp that may have resistance to viruses, (d(t)) is the number of shrimps infected from (0) to (t) die at time (t). We use this integral to express the number of shrimp die at time (t).To evaluate the performance of the model 2, we compare the simulated scenario and the biological experimental settings. Our experiments show the quantity change of dead shrimps and live shrimps with respect to time, which is consistent with the result of simulation (Supplementary Fig. 4).Mathematical model 3—use fish to control WSSWe established Model 3 for the prevention and control of WSS using fish. In Model 3, two parameters need to be determined before this model can be applied for evaluating the fish’s capability of WSS prevention and control. The two parameters are, (1) fish-feeding quantity of dead shrimp, and (2) fish-feeding ratio of dead shrimp over healthy shrimp. We obtained 1 kg grass carp’s feeding quantity of different body weights of shrimp and the feeding selectivity through experiments. The mathematical reasoning of Model 3 is as follows:To block the transmission of WSS, we apply fish to eat dead shrimp and infected shrimp. Let ({e}_{h}(t)), ({e}_{i}(t)), and ({e}_{d}(t)), respectively be the number of healthy shrimp, infected shrimp and dead shrimp eaten by fish daily at time (t), (f(t)) is the number of fish.The decrement of healthy shrimp is related to the number of infected healthy shrimp and the number of shrimp eaten by fish, as expressed in (frac{d{s}_{h}}{{dt}}=-I(t)-{e}_{h}(t)). Similarly, the dynamics of the infected shrimp is related to the number of infected healthy shrimp, the death number of infected shrimp, and the number of infected shrimp eaten by fish, as expressed in (frac{d{s}_{i}}{{dt}}=I(t)-d(t)-{e}_{i}(t)). The dynamics of dead shrimp is related to the death number of infected shrimp, and eaten by fish, as expressed in (frac{d{s}_{d}}{{dt}}=d(t)-{e}_{d}(t)). Combining the above formulae, we can write the model as follows:$$left{begin{array}{c}frac{d{s}_{h}}{{dt}}=-Ileft(tright)-{e}_{h}left(tright)quadhfill\ ,frac{d{s}_{i}}{{dt}}=Ileft(tright)-dleft(tright)-{e}_{i}left(tright)hfill\ frac{d{s}_{d}}{{dt}}=d(t)-{e}_{d}(t)hfillend{array}right.$$ (6) where ({s}_{h}left(0right)={s}_{{h}_{0}}), ({s}_{i}left(0right)={s}_{{i}_{0}}), ({s}_{d}left(0right)={s}_{{d}_{0}}) are as the initial value at (t=0). In the above model, (I(t)), (d(t)), ({e}_{h}(t)), ({e}_{i}(t)), and ({e}_{d}(t)) are respectively given as follows:$$left{begin{array}{c};, Ileft(tright)={min }left{n{s}_{d}left(tright),{s}_{h}left(tright)-alpha {s}_{{h}_{0}}right}hfill\ ;dleft(tright)={int }_{0}^{t}{min }left{n{s}_{d}left(t-tau right),{s}_{h}left(t-tau right)-alpha {s}_{{h}_{0}}right}pleft(tau right){exp }left{{int }_{t-tau }^{t}{{{{{rm{ln}}}}}}rleft(uright){du}right}dtau hfill\ {e}_{d}left(tright)={min }left{fleft(tright)cdot mcdot beta ,{s}_{d}left(tright)+dleft(tright)right}hfill\ ,{e}_{i}left(tright)={min }left{left(fleft(tright)cdot m-{e}_{d}left(tright)right)frac{{s}_{i}left(tright)+Ileft(tright)-dleft(tright)}{{s}_{i}left(tright)+{s}_{h}left(tright)-dleft(tright)},{s}_{i}left(tright)+Ileft(tright)-dleft(tright)right}hfill\ {e}_{h}left(tright)={min }left{fleft(tright)cdot m-{e}_{d}left(tright)-{e}_{i}left(tright),{s}_{h}left(tright)-Ileft(tright)right}hfill\ ;,rleft(tright)=1-frac{{e}_{i}left(tright)}{{s}_{i}left(tright)+Ileft(tright)-dleft(tright)}hfillend{array}right.$$ (7) where, (I(t)) is the same as in Eq. (4); for (d(t)), different from Eq. (5) is that we add an exponential item ({exp }left{{int }_{t-tau }^{t}{{{{{rm{ln}}}}}}rleft(uright){du}right}) to account for the infected shrimp that may be eaten by fish during the past (t) days. As for ({e}_{d}(t)) shown in Eq. (6), (m) is for that each fish eats (m) shrimps while (beta) accounts for a percentage of dead shrimp in (m) shrimp. In ({e}_{i}(t)), we introduce (frac{{s}_{i}left(tright)+Ileft(tright)-dleft(tright)}{{s}_{i}left(tright)+{s}_{h}left(tright)-dleft(tright)}) for the percentage of infected shrimp in live shrimp. ({e}_{h}(t)) accounts for the number of healthy shrimp eaten by fish. (r(t)) represents the percentage of infected shrimp not being eaten by fish. We performed the effects of 1 kg grass carps on shrimp with four different body weights. The simulated data agreed with the experimental results (Fig. 2c).The relationship among the bodyweight of one initial WSSV-infected shrimp, number of deaths, and death time distributionThree groups of 430 shrimp with a bodyweight of 1.98 ± 0.03, 6.13 ± 0.16, and 7.95 ± 0.13 g, respectively, were used. In each group, 30 shrimp were randomly selected and subjected to a two-step WSSV PCR assay. All the tested shrimp showed negative in the assay. The remaining 400 shrimps were divided equally and introduced to three experimental and one control ponds. All 12 aquariums (220 cm × 60 cm × 80 cm) were set up with a water volume of 0.5 m3 and a salinity of 8‰. Shrimp were quarantined for seven days before the experiment started. One piece of dead WSSV-infected shrimp was then introduced to each of the experimental aquariums. In addition, one piece of frozen dead shrimp (WSSV-free) was introduced to the control aquarium. Shrimp were fed once a day with artificial feed that is 2% of their body weight. Shrimp feces were timely removed, and 50% of the water in the aquarium was exchanged every day. To prevent healthy shrimps from eating the moribund shrimp but not the initial dead WSSV-infected shrimp, shrimp were observed every 10 min to identify and remove moribund shrimp from the second day of the experiment. Moribund shrimp were identified as the ones having pleopod activity, but no response to glass rod agitation. The experiment was continued until three days after the appearance of the last moribund shrimp in each aquarium. Five pieces each of moribund and survived shrimps in each aquarium were subjected to a one-step WSSV PCR assay. All moribund shrimps showed WSSV-positive, while survived shrimps showed WSSV-negative. A mathematical model (Model 1) describing the relationship among the bodyweight of one initial WSSV-infected shrimp, number of deaths, and death time distribution was established based on the experimental results.The dynamic changes of live, infected, and dead shrimps during WSSV transmissionTo determine the changes in numbers of live and dead shrimp during WSSV transmission, 9 cement ponds (315 cm × 315 cm × 120 cm) were set up with a water volume of 5 m3 and salinity of 8‰. Regarding the stocking quantity of 7.5 × 105/ha in shrimp farming production, 750 healthy shrimp with an average body weight of 7.9 g were cultured in each of the nine ponds.To prepare the WSSV acute-infected shrimp, healthy shrimp were starved for 3 days, and then fed with parts of dead WSSV-infected shrimp that are 20% of their body weights twice a day. Five shrimp were randomly selected and subjected to a one-step WSSV PCR assay. If the tested shrimp showed WSSV positive in the assay. The rest of the shrimp in the aquarium was used as the WSSV acute-infected shrimp in the following experiments.Healthy shrimp were quarantined for seven days before the experiment started. Thirty WSSV acute-infected shrimp were then introduced in each pond. Shrimps were fed once a day with artificial feed that is 2% of their body weight. The numbers of survived shrimp were counted in three ponds on the 2nd, 4th, 8th day after WSSV infection, respectively. Five dead shrimps in each pond were subjected to a one-step WSSV PCR assay, showing WSSV-positive. Based on model 1, we established a mathematical model (Model 2) to describe the dynamic changes of healthy, infected, and dead shrimps during WSSV transmission.The dead shrimp ingestion rate of fishTo determine the dead shrimp ingestion rate of grass carp (Ctenopharyngodon idellus). Three cement ponds (315 cm × 315 cm × 120 cm) were set up with a water volume of 5 m3 and a salinity of 5‰. Three grass carps with an average body weight of 0.5 kg, 1 kg, and 1.5 kg were released in each of the three ponds, respectively. The fish were raised for four days and then fed with dead shrimps with an average weight of 5.3 g. In addition, to determine the dead shrimp ingestion rate of African sharptooth catfish (Clarias gariepinus). Four cement ponds (315 cm × 315 cm × 120 cm) were set up with a water volume of 5 m3 and salinity of 3‰. One African sharptooth catfish with bodyweight of 0.262, 0.496, 0.731, and 1.502 kg was released in each of the four ponds, respectively. The fish were raised for four days and then fed with dead shrimps with an average body weight of 6.2 g. Finally, to determine the dead shrimp ingestion rate of red drum (Sciaenops ocellatus). Three cement ponds (315 cm × 315 cm × 120 cm) were set up with water volume of 5 m3 and a salinity of 5‰. One red drum with a bodyweight of 0.590, 0.654, and 0.732 kg was released in each of the three ponds, respectively. The fish were raised for four days and then fed with dead shrimps with an average body weight of 3.9 g.During the five days of the experiment, dead shrimp that were not ingested by fish were exchanged with new dead shrimps every day. Additionally, the total body weight of dead shrimp ingested by fishes was calculated by subtracting the total body weight of dead shrimp that remained in the pond from the total body weight of dead shrimp put in the pond. The shrimp ingestion rate of fish is quantified by the daily ingestion rate (total body weight of ingested shrimps per day/total body weight of fishes). The daily ingestion rate of fish was calculated for 5 days.The healthy shrimp ingestion rate of fishTo determine the healthy shrimp ingestion rate of grass carp, three experimental ponds and one control pond (315 cm × 315 cm × 120 cm) were set up with a water volume of 5 m3 and salinity of 5‰. In total, 750 healthy shrimp with an average body weight of 5.3 g were cultured in each pond. One grass carp weighting 0.956, 1.013, and 1.050 kg was released in each of the experiment ponds, respectively. No fish was released in the control pond. Every two days, 50% of the water in each pond was changed. Live shrimp that remained in each pond were counted and weighted after 10 days of the experiment.To determine the healthy shrimp ingestion rate of African sharptooth catfish, one experimental pond and one control pond (315 cm × 315 cm × 120 cm) were set up with a water volume of 5 m3 and salinity of 3‰. In total, 750 healthy shrimp with an average body weight of 2.2 g were cultured in each pond. One African sharptooth fish weighting 1.050 kg was released in the experiment pond. No fish was released in the control pond. Every 2 days, 50% of the water in each pond was changed. Live shrimp that remained in each pond were counted and weighted after 10 days of the experiment.To determine the healthy shrimp ingestion rate of red drum, three experimental ponds and one control pond (315 cm × 315 cm × 120 cm) were set up with a water volume of 5 m3 and salinity of 5‰. In total, 750 healthy shrimp with an average body weight of 2.7 g were introduced in each pond. One red drum weighting 0.519, 0.554, and 0.595 kg was released in each of the experiment ponds, respectively. No fish was released in the control pond. Every two days, 50% of the water in each pond was changed. Live shrimp that remained in each pond were counted and weighted after 10 days of the experiment.The feeding selectivity of fish on dead, infected, and healthy shrimpsTo determine the feeding selectivity of grass carp on dead, infected, and healthy shrimp, one aquarium (220 cm × 60 cm × 80 cm) was set up with a water volume of 0.5 m3 and a salinity of 5‰. Grass carp weighting 1.58 kg was cultured in the aquarium for four days before the experiment started. The diseased shrimp infected with WSSV died within two days, which makes it hard to distinguish the initial dead shrimp from the ones that were died from diseased shrimp. The diseased shrimp had reduced activity, and the activity of shrimp was reduced after the endopods and exopods were removed. Thus, shrimp with endopods and exopods removed were utilized to resemble WSSV-infected shrimp. Thirty pieces each of dead, WSSV-infected (endopods and exopods removed), and healthy shrimps were introduced in the aquarium. The mean weight of shrimp used in the experiment is 3.5 g.To determine the feeding selectivity of African sharptooth catfish on dead, infected, and healthy shrimps, one aquarium (220 cm × 60 cm × 80 cm) was set up with a water volume of 0.5 m3 and salinity of 3‰. African sharptooth catfish with body weight of 1.03 kg was cultured in the aquarium for four days before the experiment started. Thirty pieces each of dead, WSSV-infected (endopods and exopods removed), and healthy shrimps were introduced in the aquarium. The mean weight of shrimps used in the experiment is 8.4 g.During the 9 days of the experiment, the dead, infected (endopods and exopods removed), and healthy shrimp that remained in the aquarium were counted and weighed every day. New shrimps were added to ensure there are 30 pieces each of dead, infected (endopods and exopods removed), and healthy shrimp in the aquarium. The daily total body weight of shrimp that were ingested by fish in each pond was calculated by subtracting the total body weight of shrimp that remained in the pond from the total weight of shrimp put in the pond. The shrimp ingestion rate of fish is quantified by the daily ingestion rate (total body weight of ingested shrimp per day/bodyweight of fish).The suitable bodyweight of grass carp for controlling WSSTo determine the suitable bodyweight of grass carp for controlling WSS, four experimental groups and two control groups were set up. Each group consisted of three ponds (315 cm × 315 cm × 120 cm). In total, 600 healthy and 3 WSSV-infected shrimp with an average body weight of 5 g were cultured in each pond of experimental groups. One grass carp with a bodyweight of 0.3, 0.5, 1.0, 1.5 kg was released in the ponds of each experimental group, respectively. In the positive control group, 600 healthy and 3 WSSV-infected shrimp with an average body weight of 5.0 g were cultured in each of the three ponds without introducing grass carp. In the negative control group, 600 healthy shrimp with an average body weight of 5.0 g were cocultured with one grass carp weighting 1.0 kg in each of the three ponds. The numbers of live shrimp were counted after ten days of the experiment. If there were dead shrimp in the ponds, they were subjected to a one-step WSSV PCR assay. All dead shrimps showed positive for WSSV infection.The suitable bodyweight of African sharptooth catfish for controlling WSSTo determine the suitable bodyweight of African sharptooth catfish for controlling WSS, four experimental groups and two control groups were set up. Each group consisted of three ponds (315 cm × 315 cm × 120 cm). In total, 600 healthy and WSSV carrying shrimp and 3 WSSV-infected shrimp with an average body weight of 1.5 g were cultured in each pond of experimental groups. The WSSV carrying shrimp were determined as the ones that showed positive in a two-step WSSV assay. One African sharptooth catfish with a bodyweight of 0.25, 0.5, 0.75, 1.5 kg was released in the ponds of each experimental group, respectively. In the positive control group, 600 healthy and 3 WSSV-infected shrimp with an average body weight of 1.5 g were cultured in each of the three ponds without introducing African sharptooth catfish. In the negative control group, 600 healthy shrimps with an average body weight of 1.5 g were cocultured with one African sharptooth catfish weighting 1.0 kg in each pond. The numbers of live shrimp were counted after ten days of the experiment. If there were dead shrimps in the ponds, they were subjected to a one-step WSSV PCR assay. All dead shrimps showed positive for WSSV infection.The capacity of grass carp for controlling WSSTo determine the capacity of grass carp for controlling WSS, the number of WSSV-infected shrimp that could be ingested by one grass carp weighting 1 kg was evaluated. Four groups of shrimp with different body weights (1.3 ± 0.1, 2.5 ± 0.2, 5.0 ± 0.3, 7.8 ± 0.5 g) were cocultured with 1-kg grass carp in the ponds.In 1.3 ± 0.1 g group, 750 healthy shrimp were cultured in each of the nine cement ponds (315 cm × 315 cm × 120 cm). Healthy shrimps were cultured with 3, 6, 9, 12, 15, 18, and 21 pieces of WSSV-infected shrimp in each of the seven experimental ponds, respectively. One grass carp weighting 1 kg was released in each of the seven ponds. Healthy shrimp were cultured with 3 WSSV-infected shrimps in one pond as a positive control. Additionally, healthy shrimps were cultured without WSSV-infected shrimp nor grass carp in one pond as a negative control. In 2.5 ± 0.2 g group, 750 healthy shrimp were cultured with 10, 20, 30, 40, 50, 60, and 70 pieces of WSSV-infected shrimp in each of the seven experimental ponds, respectively. One grass carp weighting 1 kg was released in each of the seven ponds. Healthy shrimp were cultured with 10 WSSV-infected shrimp in one pond as a positive control. Additionally, healthy shrimps were cultured without WSSV-infected shrimp nor grass carp in one pond as a negative control. In 5.0 ± 0.3 g group, 750 healthy shrimp were cultivated with 50, 70, 90, 110, 120, 130, and 140 pieces of WSSV-infected shrimp in each of the seven experimental ponds, respectively. One grass carp weighting 1 kg was released in each of the seven ponds. Healthy shrimp were cultured with 50 WSSV-infected shrimps in one pond as a positive control. Additionally, healthy shrimps were cultured without WSSV-infected shrimps nor grass carp in one pond as a negative control. In 7.8 ± 0.5 g group, 750 healthy shrimp were cultured with 30, 40, 50, or 60 pieces of WSSV-infected shrimps in four experimental ponds, respectively. One grass carp weighting 1 kg was released in each of the four ponds. Healthy shrimp were cultured with 30 WSSV-infected shrimps in one pond as a positive control. In addition, healthy shrimps were cultured without WSSV-infected shrimps nor grass carp in one pond as a negative control.In all the ponds, shrimp were fed with artificial feed that is 2% of their body weight. And 50% of the water was changed every day. The numbers of the remaining live shrimp were counted after 15 days of the experiment. A mathematical model (Model 3) was established based on the relationship of healthy shrimp, infected shrimp, dead shrimp, and fish.Determine the numbers of grass carp and African sharptooth catfish required for controlling WSS in L. vanmamei cultivationThe number of grass carp required for controlling WSS in shrimp production was determined in Pinggang Aquaculture Base, Yangjiang, China in 2010. Forty ponds (0.34 ± 0.04 ha/pond) were divided into eight groups; each group consisted of 5 ponds. We cultured 675,000/ha of shrimp in the ponds. Shrimp were cultured for 20 days before 45, 150, 225, 300, 450, 600, 750/ha of grass carp with an average body weight of 1.0 kg were released in the ponds of group 2 to group 8. Shrimp were cultured without fish in the ponds of group 1. These 40 ponds were managed by using the same farming method. If the WSS outbreak occurred, shrimps were harvested immediately; if not, shrimps were harvested after 110 days of cultivation.The number of African sharptooth catfish required for controlling WSS in shrimp production was determined in Pinggang Aquaculture Base, Yangjiang, China in 2010. Thirty-five ponds (0.37 ± 0.06 ha/pond) were divided into seven groups; each group consisted of 5 ponds. We cultured 675,000/ha of shrimp in the ponds. Shrimp were cultured for 10 days before 150, 300, 450, 600, 750, 900/ha of African sharptooth catfish with an average body weight of 1.0 kg were released in the ponds of group 2 to group 7. Shrimp were cultured without fish in the ponds of group 1. These 35 ponds were managed by using the same farming method. If the WSS outbreak occurred, shrimps were harvested immediately; if not, shrimps were harvested after 110 days of cultivation.Validation of coculturing shrimp and grass carp for controlling WSS in L. vanmamei farmingIn 2011, the polyculture system of coculturing L. vanmamei and grass carps was validated at a farm in Maoming, Guangdong Province, China (Farm 1). Forty-six farm ponds (17.33 ha) were divided into zone A and zone B. Zone A consisted of 18 ponds with a total area of 6.03 ha, and zone B consisted of 28 ponds with a total area of 11.30 ha. The stocking quantity of shrimp in the ponds of zone A is 900,000/ha. Shrimp were cultured in the ponds for 20 days before releasing grass carps with an average body weight of 1.0 kg. The stocking quantity of fish is 317–450/ha. Shrimp were cultured without fish in the ponds of zone B, and the stocking quantity of shrimp is 900,000/ha. In 2012, we switched zones A and B, cultivating shrimp with grass carp in zone B but without fish in zone A. The stocking quantities of shrimp and fish were the same as in 2011. If a WSS outbreak occurred, shrimps were harvested immediately; if not, shrimps were harvested after 110 days of cultivation, and yields were measured.Validation of coculturing shrimp and African sharptooth catfish for controlling WSS in L. vanmamei farmingIn 2011, the polyculture system of coculturing L. vanmamei and African sharptooth catfish was validated at a farm in Qinzhou, Guangxi Province, China (Farm 2). Ninety-five farm ponds (88.2 ha) were divided into zone A and zone B. Zone A consisted of 38 ponds with a total area of 21.2 ha, and zone B consisted of 57 ponds with a total area of 67.0 ha. The stocking quantity of shrimp in the ponds of zone A is 750,000/ha. Shrimp were cultured in the ponds for 10 days before releasing African sharptooth catfish with an average body weight of 0.5 kg. The stocking quantity of fish is 525–750/ha. Shrimp were cultured without fish in the ponds of zone B, and the stocking quantity of shrimp is 750,000/ha. In 2012, we split zone B into zones B1 and B2. Shrimp were cultivated with catfish in 38 ponds of zone A and 25 ponds (27.00 ha) of zone B1, while shrimp were cultivated without fish in 32 ponds (40.00 ha) of zone B2. The stocking quantities of shrimp and fish were the same as in 2011. If WSS outbreak occurred, shrimps were harvested immediately; if not, shrimps were harvested after 110 days of cultivation, and yields were measured.Long-term validation of coculturing shrimp and fish for controlling WSS in L. vanmamei cultivationWe tested the effectiveness of using fish for controlling WSS in shrimp production at a farm in Maoming, Guangdong Province, China (Farm 1) from 2013 to 2019. In 2013, shrimp were co-cultured with African sharptooth catfish of body weight ranging from 0.5 to 0.6 kg in 13 ponds (3.73 ha). The stocking quantity of shrimp in these ponds ranges from 878,788/ha to 1,230,769/ha. And shrimp were co-cultured with grass carp of body weight ranges from 0.7 kg to 1.0 kg and African sharptooth catfish of body weight ranges from 0.5 kg to 0.6 kg in 10 ponds (3.7 ha). The stocking quantity of shrimp in these ponds ranges from 909,091/ha to 1,212,121/ha. Additionally, shrimp were cultured without fish in 11 ponds (3.63 ha). The stocking quantity of shrimp in these ponds ranges from 878,788/ha to 969,697/ha. If WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 110 days of cultivation.In 2014, shrimp were cocultured with grass carp of body weight ranging from 0.7 to 1.0 kg in 8 ponds (2.76 ha). The stocking quantity of shrimp in these ponds ranges from 833,333/ha to 1,060,606/ha. And shrimp were co-cultured with grass carp of body weight ranges from 0.7 to 1.0 kg and African sharptooth catfish of body weight ranges from 0.5 to 0.6 kg in 12 ponds (4.03 ha). The stocking quantity of shrimp in these ponds ranges from 825,000/ha to 1,060,606/ha. Additionally, shrimp were cultured without fish in 5 ponds. The stocking quantity of shrimp in these ponds was 1,060,606/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 110 days of cultivation.In 2015, shrimp were co-cultured with grass carp of body weight ranging from 0.7 to 1.0 kg in 19 ponds (7.4 ha). The stocking quantity of shrimp in these ponds ranges from 746,269 to 1,538,462/ha. In addition, shrimp were cultured without fish in 10 ponds (3.8 ha). The stocking quantity of shrimp in these ponds ranges from 750,000 to 909,091/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 110 days of cultivation.In 2016, shrimp were co-cultured with grass carp of body weight ranging from 0.7 to 1.0 kg in 19 ponds (8.11 ha). The stocking quantity of shrimp in these ponds ranges from 488,372/ha to 636,364/ha. Additionally, shrimp were cultured without fish in 8 ponds (2.84 ha). The stocking quantity of shrimp in these ponds ranges from 543,478/ha to 636,364/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 110 days of cultivation.In 2017, shrimp were cocultured with grass carp of body weight ranging from 0.7 to 1.0 kg in 6 ponds (1.56 ha). The stocking quantity of shrimp in these ponds was 961,538/ha. And shrimps were co-cultured with grass carp of body weight ranging from 0.7 kg to 1.0 kg and African sharptooth catfish of body weight ranges from 0.5 to 0.6 kg in 12 ponds (3.96 ha). The stocking quantity of shrimp in these ponds ranges from 848,485/ha to 909,091/ha. Additionally, shrimp were cultured without fish in 9 ponds (2.76 ha). The stocking quantity of shrimp in these ponds ranges from 848,485/ha to 961,538/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 110 days of cultivation.In 2018, shrimp were cocultured with grass carp of body weight ranging from 0.7g to 1.0 kg in 22 ponds (9.24 ha). The stocking quantity of shrimp in these ponds ranges from 454,545/ha to 869,565/ha. Additionally, shrimp were cultured without fish in 9 ponds (3.36 ha). The stocking quantity of shrimp in these ponds ranges from 695,652/ha to 861,111/ha. If a WSS outbreak occurred, shrimp were harvested; if not, shrimp were harvested after 110 days of cultivation.In 2019, shrimp were cocultured with grass carp of body weight ranging from 0.7 to 1.0 kg in 30 ponds (11.31 ha). The stocking quantity of shrimp in these ponds ranges from 652,174/ha to 1,000,000/ha. Additionally, shrimp were cultured without fish in 10 ponds (3.57 ha). The stocking quantity of shrimp in these ponds ranges from 666,667/ha to 1,000,000/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 110 days of cultivation.Validation of coculturing shrimp and brown-marbled grouper for controlling WSS in P. monodon farmingIn 2013, the polyculture system of coculturing P. monodon and brown-marbled grouper was validated at a farm in Changjiang, Hainan Province, China (Farm 3). We cultured 6 × 105/ha of non-SPF shrimp in 6 ponds (1.60 ha) for 30 days and then introduced 600~750/ha of brown-marbled grouper with an average body weight of 0.1 kg. Shrimp were also cultured without fish in 3 ponds (0.8 ha). The stocking quantity of shrimp in these ponds is 6 × 105/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 150 days of cultivation, and yields were measured.In 2014, we cultured 6 × 105/ha of non-SPF shrimp in 6 ponds (1.60 ha) for 30 days and then introduced 600–750/ha of brown-marbled grouper with an average body weight of 0.1 kg. Shrimps were also cultured without fish in 3 ponds (0.8 ha). The stocking quantity of shrimp in these ponds is 6 × 105/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 150 days of cultivation, and yields were measured.Validation of coculturing shrimp and branded gobies for controlling WSS in M. japonica farmingIn 2013, the polyculture system of coculturing M. japonica and branded gobies was validated at a farm in Qingdao, Shandong Province, China (Farm 4). We cultured 1.5 × 105/ha of non-SPF shrimp in 10 ponds (13.40 ha) for 30 days and then introduced 750~900/ha of branded gobies with an average body weight of 0.05 kg. Shrimp were also cultured without fish in 5 ponds (6.70 ha). The stocking quantity of shrimp in these ponds is 1.5 × 105/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 100 days of cultivation, and yields were measured.In 2014, we cultured 1.5 × 105/ha of non-SPF shrimp in 10 ponds (13.40 ha) for 30 days and then introduced 750~900/ha of branded gobies with an average body weight of 0.1 kg. Shrimp were also cultured without fish in 5 ponds (6.70 ha). The stocking quantity of shrimp in these ponds is 1.5 × 105/ha. If a WSS outbreak occurred, shrimp were harvested immediately; if not, shrimp were harvested after 100 days of cultivation, and yields were measured.Promotion of the polyculture system at a farmers’ association in Nansha, ChinaWhen we promoted the polyculture system at the farmers’ association in 2015, only 6 farmers decided to adopt the system, as most of the farmers worried that fish would ingest healthy shrimp. Each of the 6 farmers introduced 225,000, 360,000, and 360,000 P.monodon postlarva to his/her earthen pond (3 ha) on March 28, May 8, and June 15, respectively. And 1350 grass carps with an average body weight of 1 kg were released in the ponds on April 30. These farmers harvested shrimp from May to November, and grass carp on December 14. The yields of shrimp and fish of these six ponds were recoded. The other farmers in the association introduced 225,000 and 360,000 of P.monodon postlarva to their ponds (3 ha) on March 28 and May 8, respectively. WSS outbreaks occur in their ponds from May 15 to May 23. Therefore, these farmers only harvested shrimp in May. Six ponds were randomly selected, and the yields of these ponds were recorded.Promotion of the polyculture system at a farmers’ association in Tanghai, ChinaFarmers at the farmers’ association used to culture 1500/ha of F. chinensis in earthen pond (5 ha) before the promotion of the polyculture system in 2015. The yields of 10 randomly selected ponds in 2014 were recorded. In 2015, farmers at the association started to culture 8,000/ha of F. chinensis in their ponds. The shrimp were cultured 20 days before 800/ha of branded gobies with an average body weight of 0.05 kg were released in the ponds. Branded gobies were cultivated for 15 days before introducing to the ponds. Shrimps were harvested after 120 days of cultivation. The yields of ten randomly selected ponds were recorded.Statistics and reproducibilityAlpha levels of 0.05 were regarded as statistically significant throughout the study. Three replicates were set up for each experiment to confirm the reproducibility of the data. All data are reported as the mean ± standard errors.Reporting summaryFurther information on research design is available in the Nature Research Reporting Summary linked to this article. More

1.Brazil. USDA Foreign Agricultural Service https://www.fas.usda.gov/regions/brazil (2019).2.Planilha do PIB do Agronegócio Brasileiro de 1996 a 2018 (Centro de Estudos Avançados em Economia Aplicada, 2018); https://www.cepea.esalq.usp.br/br/pib-do-agronegocio-brasileiro.aspx3.Boletim da Safra de Grãos. Companhia Nacional de Abastecimento https://www.conab.gov.br/info-agro/safras/graos/boletim-da-safra-de-graos (2020).4.Projeções do Agronegócio: Brasil 2017/18 a 2027/28 Projeções de Longo Prazo (Ministério da Agricultura, Pecuária e Abastecimento, 2018).5.Atlas Irrigação: Uso da Água na Agricultura Irrigada (Agência Nacional de Águas, 2017).6.Costa, M. H. et al. Climate risks to Amazon agriculture suggest a rationale to conserve local ecosystems. Front. Ecol. Environ. 17, 584–590 (2019).
Google Scholar 
7.Fu, R. et al. Increased dry-season length over southern Amazonia in recent decades and its implication for future climate projection. Proc. Natl Acad. Sci. USA 110, 18110–18115 (2013).CAS 

Google Scholar 
8.Spera, S. A., Galford, G. L., Coe, M. T., Macedo, M. N. & Mustard, J. F. Land-use change affects water recycling in Brazil’s last agricultural frontier. Glob. Change Biol. 22, 3405–3413 (2016).
Google Scholar 
9.Abrahão, G. M. & Costa, M. H. Evolution of rain and photoperiod limitations on the soybean growing season in Brazil: the rise (and possible fall) of double-cropping systems. Agric. Meteorol. 256–257, 32–45 (2018).
Google Scholar 
10.Silvério, D. V. et al. Agricultural expansion dominates climate changes in southeastern Amazonia: the overlooked non-GHG forcing. Environ. Res. Lett. 10, 104015 (2015).
Google Scholar 
11.Barkhordarian, A., Saatchi, S. S., Behrangi, A., Loikith, P. C. & Mechoso, C. R. A recent systematic increase in vapor pressure deficit over tropical South America. Sci. Rep. 9, 15331 (2019).
Google Scholar 
12.Barkhordarian, A., von Storch, H., Zorita, E., Loikith, P. C. & Mechoso, C. R. Observed warming over northern South America has an anthropogenic origin. Clim. Dyn. 51, 1901–1914 (2018).
Google Scholar 
13.Leite‐Filho, A. T., Costa, M. H. & Fu, R. The southern Amazon rainy season: the role of deforestation and its interactions with large‐scale mechanisms. Int. J. Climatol. 40, 2328–2341 (2020).
Google Scholar 
14.FAOSTAT (Food and Agriculture Organization of the United Nations, 2020); http://www.fao.org/faostat/en/#data/QC15.Presidência da República Secretaria-Geral Subchefia para Assuntos Jurídicos (Ministério da Agricultura, 2015).16.Rashid, M. A. et al. Impact of heat-wave at high and low VPD on photosynthetic components of wheat and their recovery. Environ. Exp. Bot. 147, 138–146 (2018).CAS 

Google Scholar 
17.Lobell, D. B. et al. Greater sensitivity to drought accompanies maize yield increase in the U.S. Midwest. Science 344, 516–519 (2014).CAS 

Google Scholar 
18.Fletcher, A. L., Sinclair, T. R. & Allen, L. H. Transpiration responses to vapor pressure deficit in well watered ‘slow-wilting’ and commercial soybean. Environ. Exp. Bot. 61, 145–151 (2007).CAS 

Google Scholar 
19.Bunce, J. A. Comparative responses of leaf conductance to humidity in single attached leaves. J. Exp. Bot. 32, 629–634 (1981).
Google Scholar 
20.Kiniry, J. et al. Radiation-use efficiency response to vapor pressure deficit for maize and sorghum. Field Crops Res. 56, 265–270 (1998).
Google Scholar 
21.Spera, S. A. et al. Recent cropping frequency, expansion, and abandonment in Mato Grosso, Brazil had selective land characteristics. Environ. Res. Lett. 9, 064010 (2014).
Google Scholar 
22.Dias, L. C. P., Pimenta, F. M., Santos, A. B., Costa, M. H. & Ladle, R. J. Patterns of land use, extensification, and intensification of Brazilian agriculture. Glob. Change Biol. 22, 2887–2903 (2016).
Google Scholar 
23.Cohn, A. S., Vanwey, L. K., Spera, S. A. & Mustard, J. F. Cropping frequency and area response to climate variability can exceed yield response. Nat. Clim. Change 6, 601–604 (2016).
Google Scholar 
24.Morton, D. C. et al. Reevaluating suitability estimates based on dynamics of cropland expansion in the Brazilian Amazon. Glob. Environ. Change 37, 92–101 (2016).
Google Scholar 
25.Duursma, R. A. et al. The peaked response of transpiration rate to vapour pressure deficit in field conditions can be explained by the temperature optimum of photosynthesis. Agric. Meteorol. 189–190, 2–10 (2014).
Google Scholar 
26.Spera, S. A., Winter, J. M. & Partridge, T. F. Brazilian maize yields negatively affected by climate after land clearing. Nat. Sustain. 3, 845–852 (2020).
Google Scholar 
27.Cirino, P. H., Féres, J. G., Braga, M. J. & Reis, E. Assessing the impacts of ENSO-related weather effects on the Brazilian agriculture. Proc. Econ. Financ. 24, 146–155 (2015).
Google Scholar 
28.Pereira, P. A. A., Martha, G. B., Santana, C. A. & Alves, E. The development of Brazilian agriculture: future technological challenges and opportunities. Agric. Food Secur. 1, 4 (2012).
Google Scholar 
29.Marengo, J. A. & Bernasconi, M. Regional differences in aridity/drought conditions over Northeast Brazil: present state and future projections. Climatic Change 129, 103–115 (2015).
Google Scholar 
30.Naylor, R. L. Energy and resource constraints on intensive agricultural production. Annu. Rev. Energy Environ. 21, 99–123 (1996).
Google Scholar 
31.Getirana, A. Extreme water deficit in Brazil detected from space. J. Hydrometeorol. 17, 591–599 (2016).
Google Scholar 
32.Lathuillière, M. J., Coe, M. T. & Johnson, M. S. A review of green- and blue-water resources and their trade-offs for future agricultural production in the Amazon Basin: what could irrigated agriculture mean for Amazonia? Hydrol. Earth Syst. Sci. 20, 2179–2194 (2016).
Google Scholar 
33.Dobrovolski, R. & Rattis, L. Water collapse in Brazil: the danger of relying on what you neglect. Nat. Conserv. 13, 80–83 (2015).
Google Scholar 
34.da Silva, A. L. et al. Water appropriation on the agricultural frontier in western Bahia and its contribution to streamflow reduction: revisiting the debate in the Brazilian Cerrado. Water 13, 1054 (2021).
Google Scholar 
35.Pousa, R. et al. Climate change and intense irrigation growth in western Bahia, Brazil: the urgent need for hydroclimatic monitoring. Water 11, 933 (2019).
Google Scholar 
36.Ort, D. R. & Long, S. P. Limits on yields in the corn belt. Science 344, 484–485 (2014).CAS 

Google Scholar 
37.de Bossoreille de Ribou, S., Douam, F., Hamant, O., Frohlich, M. W. & Negrutiu, I. Plant science and agricultural productivity: why are we hitting the yield ceiling? Plant Sci. 210, 159–176 (2013).
Google Scholar 
38.Long, S. P. & Ort, D. R. More than taking the heat: crops and global change. Curr. Opin. Plant Biol. 13, 240–247 (2010).
Google Scholar 
39.Pommer, C. V. & Barbosa, W. The impact of breeding on fruit production in warm climates of Brazil. Rev. Bras. Frutic. 31, 612–634 (2009).
Google Scholar 
40.Lenka, N. K. et al. Carbon dioxide and temperature elevation effects on crop evapotranspiration and water use efficiency in soybean as affected by different nitrogen levels. Agric. Water Manag. 230, 105936 (2020).
Google Scholar 
41.Soares, W. R., Marengo, J. A. & Nobre, C. A. Assessment of warming projections and probabilities for Brazil in Climate Change Risks in Brazil (eds Nobre, C. et al.) 7–30 (Springer, 2019); https://doi.org/10.1007/978-3-319-92881-4_242.Schwalm, C. R., Glendon, S. & Duffy, P. B. RCP8.5 tracks cumulative CO2 emissions. Proc. Natl Acad. Sci. USA 117, 19656–19657 (2020).CAS 

Google Scholar 
43.Schwalm, C. R., Glendon, S. & Duffy, P. B. Reply to Hausfather and Peters: RCP8.5 is neither problematic nor misleading. Proc. Natl Acad. Sci. USA 117, 27793–27794 (2020).CAS 

Google Scholar 
44.Sistematização das Informações sobre Recursos Naturais—Mapa de Biomas do Brasil (Instituto Brasileiro de Geografia e Estatística, 2006); https://www.ibge.gov.br/geociencias/cartas-e-mapas/informacoes-ambientais/15842-biomas.html?=&t=downloads45.Base Cartográfica Continua Do Brasil, Escala 1:250.000—BC250 (Instituto Brasileiro de Geografia e Estatística, 2019); https://geoftp.ibge.gov.br/cartas_e_mapas/bases_cartograficas_continuas/bc250/versao2019/informacoes_tecnicas/Documentacao_bc250_v2019.pdf46.Campos, J., de, O. & Chaves, H. M. L. Tendências e variabilidades nas séries históricas de precipitação mensal e anual no bioma Cerrado no período 1977–2010. Rev. Bras. Meteorol. 35, 157–169 (2020).
Google Scholar 
47.Debortoli, N. S. et al. Rainfall patterns in the southern Amazon: a chronological perspective (1971–2010). Climatic Change 132, 251–264 (2015).
Google Scholar 
48.Oliveira, P. T. S. et al. Trends in water balance components across the Brazilian Cerrado. Water Resour. Res. 50, 7100–7114 (2014).
Google Scholar 
49.Panisset, J. S. et al. Contrasting patterns of the extreme drought episodes of 2005, 2010 and 2015 in the Amazon Basin. Int. J. Climatol. 38, 1096–1104 (2018).
Google Scholar 
50.Cai, W. et al. Increasing frequency of extreme El Niño events due to greenhouse warming. Nat. Clim. Change 4, 111–116 (2014).
Google Scholar 
51.Jiménez-Muñoz, J. C. et al. Record-breaking warming and extreme drought in the Amazon rainforest during the course of El Niño 2015–2016. Sci. Rep. 6, 33130 (2016).
Google Scholar 
52.Marengo, J. A. & Espinoza, J. C. Extreme seasonal droughts and floods in Amazonia: causes, trends and impacts. Int. J. Climatol. 36, 1033–1050 (2016).
Google Scholar 
53.Funk, C. et al. The climate hazards infrared precipitation with stations—a new environmental record for monitoring extremes. Sci. Data 2, 150066 (2015).
Google Scholar 
54.Abatzoglou, J. T., Dobrowski, S. Z., Parks, S. A. & Hegewisch, K. C. TerraClimate, a high-resolution global dataset of monthly climate and climatic water balance from 1958–2015. Sci. Data 5, 170191 (2018).
Google Scholar 
55.Harris, I., Osborn, T. J., Jones, P. & Lister, D. Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Sci. Data 7, 109 (2020).
Google Scholar 
56.Gorelick, N. et al. Google Earth Engine: planetary-scale geospatial analysis for everyone. Remote Sens. Environ. 202, 18–27 (2017).
Google Scholar 
57.R Core Team R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2019).58.Challinor, A. J. & Wheeler, T. R. Crop yield reduction in the tropics under climate change: processes and uncertainties. Agric. Meteorol. 148, 343–356 (2008).
Google Scholar 
59.Bates, D. et al. lme4. R package version (2012).60.Barton, K. MuMIn: Multi-model inference. R package version 1.0.0 (2009).61.Arvor, D., Dubreuil, V., Ronchail, J., Simões, M. & Funatsu, B. M. Spatial patterns of rainfall regimes related to levels of double cropping agriculture systems in Mato Grosso (Brazil). Int. J. Climatol. 34, 2622–2633 (2014).
Google Scholar 
62.Hengl, T. et al. SoilGrids250m: global gridded soil information based on machine learning. PLoS ONE 12, e0169748 (2017).
Google Scholar 
63.Brill, F., Passuni Pineda, S., Espichán Cuya, B. & Kreibich, H. A data-mining approach towards damage modelling for El Niño events in Peru. Geomat. Nat. Hazards Risk 11, 1966–1990 (2020).
Google Scholar 
64.Rattis, L. ludmilarattis/effect-of-climate-on–agriculture: Rattis_etal_NCC_2021. Zenodo https://zenodo.org/badge/latestdoi/271879616 (2021).65.Malhi, Y. et al. Exploring the likelihood and mechanism of a climate-change-induced dieback of the Amazon rainforest. Proc. Natl Acad. Sci. USA 106, 20610–20615 (2009).CAS 

Google Scholar 
66.Castanho, A. D. A. et al. Potential shifts in the aboveground biomass and physiognomy of a seasonally dry tropical forest in a changing climate. Environ. Res. Lett. 15, 034053 (2020).
Google Scholar 
67.Allen, R. G. et al. The ASCE Standardized Reference Evapotranspiration Equation (American Society of Civil Engineers, 2005).68.IPCC Climate Change 2014: Synthesis Report (eds Core Writing Team, Pachauri, R. K. & Meyer, L. A.) (IPCC, 2014).69.Koutroulis, A. G., Grillakis, M. G., Tsanis, I. K. & Papadimitriou, L. Evaluation of precipitation and temperature simulation performance of the CMIP3 and CMIP5 historical experiments. Clim. Dyn. 47, 1881–1898 (2016).
Google Scholar 
70.Análise Territorial para o Desenvolvimento da Agricultura Irrigada no Brasil (Ministério da Integração Nacional, 2014). More

1.O’Connor. Whale Watching Worldwide: Tourism Numbers, Expenditures and Expanding Economic Benefits. A special report from the International Fund for Animal Welfare Yarmouth, USA. Prepared by Economists at Large. www.ecolarge.com (2009). Accessed 15 Apr 2021.2.Cisneros-Montemayor, A. M., Sumaila, U. R., Kaschner, K. & Pauly, D. The global potential for whale watching. Mar. Policy 34, 1273–1278 (2010).Article 

Google Scholar 
3.Parsons, E. C. M. The negative impacts of whale-watching. J. Mar. Biol. 2012, 1–9 (2012).ADS 
Article 

Google Scholar 
4.Hoyt, E. & Hvenegaard, G. T. A review of whale-watching and whaling. Coast. Manag. 30, 381–399 (2002).Article 

Google Scholar 
5.Cunningham, P. A., Huijbens, E. H. & Wearing, S. L. From whaling to whale watching: Examining sustainability and cultural rhetoric. J. Sustain. Tour. 20, 143–161 (2012).Article 

Google Scholar 
6.Senigaglia, V. et al. Meta-analyses of whale-watching impact studies: Comparisons of cetacean responses to disturbance. Mar. Ecol. Prog. Ser. 542, 251–263 (2016).ADS 
CAS 
Article 

Google Scholar 
7.Bejder, L. et al. Decline in relative abundance of bottlenose dolphins exposed to long-term disturbance. Conserv. Biol. 20, 1791–1798 (2006).PubMed 
Article 
PubMed Central 

Google Scholar 
8.Lusseau, D., Slooten, L. & Currey, R. J. C. Unsustainable dolphin-watching tourism in Fiordland, New Zealand. Tour. Mar. Environ. 3, 173–178 (2006).Article 

Google Scholar 
9.Erbe, C. Underwater noise of whale-watching boats and potential effects on killer whales (Orcinus orca), based on an acoustic impact model. Mar. Mamm. Sci. 18, 394–418 (2002).Article 

Google Scholar 
10.Sprogis, K. R., Videsen, S. & Madsen, P. T. Vessel noise levels drive behavioural responses of humpback whales with implications for whale-watching. Elife 9, 1–17 (2020).Article 

Google Scholar 
11.Au, W. W. L. The Sonar of Dolphins (Springer, 1993).Book 

Google Scholar 
12.Tyack, P. L. & Clark, C. W. Communication and acoustic behavior of dolphins and whales. In Hearing by Whales and Dolphins (eds. Au, W. W. L., Fay, R. R. & Popper, A. N.) 156–224 (Springer, 2000). https://doi.org/10.1007/978-1-4612-1150-1_4.13.Lesage, V., Barrette, C., Kingsley, M. C. S. & Sjare, B. The effect of vessel noise on the vocal behavior of belugas in the St. Lawrence River estuary, Canada. Mar. Mamm. Sci. 15, 65–84 (1999).Article 

Google Scholar 
14.Jensen, F. H. et al. Vessel noise effects on delphinid communication. Mar. Ecol. Prog. Ser. 395, 161–175 (2009).ADS 
Article 

Google Scholar 
15.Pirotta, E. et al. Vessel noise affects beaked whale behavior: Results of a dedicated acoustic response study. PLoS One 7, e42535 (2012).ADS 
CAS 
PubMed 
PubMed Central 
Article 

Google Scholar 
16.Slabbekoorn, H. et al. A noisy spring: The impact of globally rising underwater sound levels on fish. Trends Ecol. Evol. 25, 419–427 (2010).PubMed 
Article 

Google Scholar 
17.Andrew, R. K., Howe, B. M. & Mercer, J. A. Long-time trends in ship traffic noise for four sites off the North American West Coast. J. Acoust. Soc. Am. 129, 642–651 (2011).ADS 
PubMed 
Article 

Google Scholar 
18.Miksis-Olds, J. L. & Nichols, S. M. Is low frequency ocean sound increasing globally?. J. Acoust. Soc. Am. 139, 501–511 (2016).ADS 
PubMed 
Article 

Google Scholar 
19.Payne, R. & Webb, D. Orientation by means of long range acoustic signaling in Baleen whales. Ann. N. Y. Acad. Sci. 188, 110–141 (1971).ADS 
CAS 
PubMed 
Article 

Google Scholar 
20.Melcón, M. L. et al. Blue whales respond to anthropogenic noise. PLoS One 7, e32681 (2012).ADS 
PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
21.Romagosa, M. et al. Underwater ambient noise in a baleen whale migratory habitat off the Azores. Front. Mar. Sci. 4, 1–14 (2017).Article 

Google Scholar 
22.Aguilar Soto, N. et al. Does intense ship noise disrupt foraging in deep-diving cuvier’s beaked whales (Ziphius cavirostris)?. Mar. Mamm. Sci. 22, 690–699 (2006).Article 

Google Scholar 
23.Wisniewska, D. M. et al. High rates of vessel noise disrupt foraging in wild harbour porpoises (Phocoena phocoena). Proc. R. Soc. B Biol. Sci. 285, 20172314 (2018).Article 

Google Scholar 
24.Hermannsen, L., Beedholm, K., Tougaard, J. & Madsen, P. T. High frequency components of ship noise in shallow water with a discussion of implications for harbor porpoises (Phocoena phocoena). J. Acoust. Soc. Am. 136, 1640–1653 (2014).ADS 
PubMed 
Article 

Google Scholar 
25.Wladichuk, J., Hannay, D. E., MacGillivray, A. O., Li, Z. & Thornton, S. J. Systematic source level measurements of whale watching vessels and other small boats. J. Ocean Technol. 14, 108–126 (2019).
Google Scholar 
26.Arranz, P., Aguilar de Soto, N., Madsen, P. T. & Sprogis, K. R. Whale-watch vessel noise levels with applications to whale-watching guidelines and conservation. Mar. Policy 134, 104776 (2021).Article 

Google Scholar 
27.Higham, J., Bejder, L. & Williams, R. Whale-Watching: Sustainable Tourism and Ecological Management (Cambridge University Press, 2014).Book 

Google Scholar 
28.Montero, R. & Arechavaleta, M. Distribution patterns: Relationships between depths, sea surface temperature, and habitat use of Short-finned pilot whales south-west of Tenerife. Eur. Res. Cetaceans 10, 193–198 (1996).
Google Scholar 
29.Servidio, A. et al. Site fidelity and movement patterns of short-finned pilot whales within the Canary Islands: Evidence for resident and transient populations. Aquat. Conserv. Mar. Freshw. Ecosyst. 29, 227–241 (2019).Article 

Google Scholar 
30.Würsig, B. Ethology and Behavioral Ecology of Odontocetes (Springer, 2019). https://doi.org/10.1007/978-3-030-16663-2_23.Book 

Google Scholar 
31.Sequeira, M. et al. Review of whalewatching activities in mainland Portugal, the Azores, Madeira and Canary archipelagos and the Strait of Gibraltar. J. Cetacean Res. Manag. SC61/WW11, 1–40 (2009).
32.IWC whale-watching handbook. https://wwhandbook.iwc.int/es/case-studies/canary-islands-spain#Accessed 15 Apr 2021.33.Hoyt, E. Tourism. In Encyclopedia of Marine Mammals 1010–1014 (2018). https://doi.org/10.1016/B978-0-12-804327-1.00262-4.
34.Kasuya, T. & Matsui, S. Age determination and growth of the short-finned pilot whale off the Pacific coast of Japan. Sci. Rep. Whales Res. Inst. 35, 57–91 (1984).
Google Scholar 
35.Reddy, M., Kamolnick, T., Skaar, D., Curry, C. & Ridgway, S. Bottlenose dolphins: Energy consumption during pregnancy, lactation, and growth. Int. Mar. Mammal Trainers Assoc. Conf. Proc. 30–37 (1991).36.Srinivasan, M., Swannack, T. M., Grant, W. E., Rajan, J. & Würsig, B. To feed or not to feed? Bioenergetic impacts of fear-driven behaviors in lactating dolphins. Ecol. Evol. 8, 1384–1398 (2018).PubMed 
Article 

Google Scholar 
37.Marsh, H. & Kasuya, T. Changes in the role of a female pilot whale with age. In Dolphin Societies (eds. Pryor, K. & Norris, K.S.) 281–286 (University of California Press, 1991).38.Augusto, J. F., Frasier, T. R. & Whitehead, H. Characterizing alloparental care in the pilot whale (Globicephala melas) population that summers off Cape Breton, Nova Scotia, Canada. Mar. Mamm. Sci. 33, 440–456 (2017).Article 

Google Scholar 
39.Quick, N., Scott-hayward, L., Sadykova, D., Nowacek, D. & Read, A. Effects of a scientific echo sounder on the behavior of short-finned pilot whales (Globicephala macrorhynchus). Can. J. Fish. Aquat. Sci. 74, 716–726 (2017).Article 

Google Scholar 
40.Würsig, B. & Jefferson, T. A. Methods of photo-identification for small cetaceans. Report of the International Whaling Commission (1990).41.Greenhow, D. R., Brodsky, M. C., Lingenfelser, R. G. & Mann, D. A. Hearing threshold measurements of five stranded short-finned pilot whales (Globicephala macrorhynchus). J. Acoust. Soc. Am. 135, 531–536 (2014).ADS 
PubMed 
Article 
PubMed Central 

Google Scholar 
42.Tougaard, J. & Beedholm, K. Practical implementation of auditory time and frequency weighting in marine bioacoustics. Appl. Acoust. 145, 137–143 (2019).Article 

Google Scholar 
43.Pérez, J. M., Jensen, F. H., Rojano-Doñate, L. & Aguilar de Soto, N. Different modes of acoustic communication in deep-diving short-finned pilot whales (Globicephala macrorhynchus). Mar. Mamm. Sci. 33, 59–79 (2017).Article 

Google Scholar 
44.Pacini, A. F. et al. Audiogram of a formerly stranded long-finned pilot whale (Globicephala melas) measured using auditory evoked potentials. J. Exp. Biol. 213, 3138–3143 (2010).CAS 
PubMed 
Article 

Google Scholar 
45.Christiansen, F., Rojano-Doñate, L., Madsen, P. T. & Bejder, L. Noise levels of multi-rotor unmanned aerial vehicles with implications for potential underwater impacts on marine mammals. Front. Mar. Sci. 3, 1–9 (2016).Article 

Google Scholar 
46.Christiansen, F., Nielsen, M. L. K., Charlton, C., Bejder, L. & Madsen, P. T. Southern right whales show no behavioral response to low noise levels from a nearby unmanned aerial vehicle. Mar. Mamm. Sci. 36, 953–963 (2020).Article 

Google Scholar 
47.Giles, A. B. et al. Responses of bottlenose dolphins (Tursiops spp.) to small drones. Aquat. Conserv. Mar. Freshw. Ecosyst. 31, 677–684 (2021).Article 

Google Scholar 
48.Nielsen, M. L. K., Sprogis, K. R., Bejder, L., Madsen, P. T. & Christiansen, F. Behavioural development in southern right whale calves. Mar. Ecol. Prog. Ser. 629, 219–234 (2019).ADS 
Article 

Google Scholar 
49.Hofmann, B., Scheer, M. & Behr, I. P. Underwater behaviors of short-finned pilot whales (Globicephala macrorhynchus) off Tenerife. Mammalia 68, 221–224 (2004).Article 

Google Scholar 
50.Lockyer, C. All creatures great and smaller: A study in cetacean life history energetics. J. Mar. Biol. Assoc. U. K. 87, 1035–1045 (2007).Article 

Google Scholar 
51.Hin, V., Harwood, J. & de Roos, A. M. Bio-energetic modeling of medium-sized cetaceans shows high sensitivity to disturbance in seasons of low resource supply. Ecol. Appl. 29, 1–19 (2019).Article 

Google Scholar 
52.Christiansen, F., Rasmussen, M. H. & Lusseau, D. Inferring energy expenditure from respiration rates in minke whales to measure the effects of whale watching boat interactions. J. Exp. Mar. Biol. Ecol. 459, 96–104 (2014).Article 

Google Scholar 
53.Mann, J., Connor, R. C., Tyack, P. L. & Whitehead, H. Cetacean Societies: Field Studies of Dolphins and Whales (The University of Chicago Press, 2000).
Google Scholar 
54.Zuur, A. F., Ieno, E. N. & Elphick, C. S. A protocol for data exploration to avoid common statistical problems. Methods Ecol. Evol. 1, 3–14 (2010).Article 

Google Scholar 
55.R Development Core Team. R: A Language and Environment for Statistical Computing. (2018).56.National Marine Fisheries Service. Technical guidance for assessing the effects of anthropogenic sound on marine mammal hearing underwater acoustic thresholds for onset of permanent and temporary threshold shifts. NOAA Technical Memorandum NMFS-OPR-55. (2016).57.Heimlich-Boran, J. R. Social Organisation of the Short-finned Pilot Whale, Globicephala macrorhynchus, with Special Reference to the Comparative Social Ecology of Delphinids (University of Cambridge, 1993).
Google Scholar 
58.Hastie, G. D., Wilson, B., Tufft, L. H. & Thompson, P. M. Bottlenose dolphins increase breathing synchrony in response to boat traffic. Mar. Mamm. Sci. 19, 74–084 (2003).Article 

Google Scholar 
59.Williams, R. & Noren, D. P. Swimming speed, respiration rate, and estimated cost of transport in adult killer whales. Mar. Mamm. Sci. 25, 327–350 (2009).Article 

Google Scholar 
60.Senigaglia, V. & Whitehead, H. Synchronous breathing by pilot whales. Mar. Mamm. Sci. 28, 213–219 (2012).Article 

Google Scholar 
61.Aguilar Soto, N. et al. Cheetahs of the deep sea: Deep foraging sprints in short-finned pilot whales off Tenerife (Canary Islands). J. Anim. Ecol. 77, 936–947 (2008).PubMed 
Article 

Google Scholar 
62.Ariza, A. et al. Vertical distribution, composition and migratory patterns of acoustic scattering layers in the Canary Islands. J. Mar. Syst. 157, 82–91 (2016).Article 

Google Scholar 
63.Owen, K., Andrews, R. D., Baird, R. W., Schorr, G. S. & Webster, D. L. Lunar cycles influence the diving behavior and habitat use of short-finned pilot whales around the main Hawaiian Islands. Mar. Ecol. Prog. Ser. 629, 193–206 (2019).ADS 
Article 

Google Scholar 
64.Kasuya, T. & Marsh, H. Life history and reproductive biology of the short-finned pilot whale, Globicephala macrorhynchus, off the Pacific coast of Japan. Rep. Int. Whal. Commun. Spec. Iss. 6, 259–310 (1984).
Google Scholar 
65.Barton, E. et al. The transition zone of the canary current upwelling region. Prog. Ocean. 41, 455–504 (1998).Article 

Google Scholar 
66.Olson, P. A. Pilot Whales: Globicephala melas and G. macrorhynchus. In Encyclopedia of Marine Mammals, 3rd ed (eds. Würsig, B., Thewissen, J. G. M. & Kovacs, K. M.) 701–705 (Academic Press, 2018). https://doi.org/10.1016/B978-0-12-804327-1.00194-1.67.Puig-Lozano, R. et al. Retrospective study of fishery interactions in stranded cetaceans, Canary Islands. Front. Vet. Sci. 7, 1–15 (2020).Article 

Google Scholar 
68.Almunia, J., Delponti, P. & Rosa, F. Using automatic identification system (AIS) data to estimate whale watching effort. Front. Mar. Sci. 8, 827 (2021).Article 

Google Scholar 
69.Schlundt, C. E. et al. Auditory evoked potentials in two short-finned pilot whales (Globicephala macrorhynchus). J. Acoust. Soc. Am. 129, 1111–1116 (2011).ADS 
PubMed 
Article 

Google Scholar 
70.Visser, F. et al. Risso’s dolphins alter daily resting pattern in response to whale watching at the Azores. Mar. Mamm. Sci. 27, 366–381 (2011).Article 

Google Scholar 
71.Constantine, R., Brunton, D. H. & Dennis, T. Dolphin-watching tour boats change bottlenose dolphin (Tursiops truncatus) behaviour. Biol. Conserv. 117, 299–307 (2004).Article 

Google Scholar 
72.Stensland, E. & Berggren, P. Behavioural changes in female Indo-Pacific bottlenose dolphins in response to boat-based tourism. Mar. Ecol. Prog. Ser. 332, 225–234 (2007).ADS 
Article 

Google Scholar 
73.Jensen, F. H., Perez, J. M., Johnson, M., Soto, N. A. & Madsen, P. T. Calling under pressure: Short-finned pilot whales make social calls during deep foraging dives. Proc. R. Soc. B Biol. Sci. 278, 3017–3025 (2011).Article 

Google Scholar 
74.Kiszka, J. J., Caputo, M., Méndez-Fernandez, P. & Fielding, R. Feeding ecology of elusive Caribbean killer whales inferred from bayesian stable isotope mixing models and whalers’ ecological knowledge. Front. Mar. Sci. 8, 1–11 (2021).Article 

Google Scholar 
75.Whitehead, H. Babysitting, dive synchrony, and indications of alloparental care in sperm whales. Behav. Ecol. Sociobiol. 38, 237–244 (1996).Article 

Google Scholar 
76.Konrad, C. M. Kinship in Sperm Whale Society: Effects on Association, Alloparental Care and Vocalization (Dalhousie University, 2017).
Google Scholar 
77.Leung, E. S., Vergara, V. & Barrett-Lennard, L. G. Allonursing in captive belugas (Delphinapterus leucas). Zoo Biol. 29, 633–637 (2010).PubMed 
Article 

Google Scholar 
78.Houston, A. I. & Carbone, C. The optimal allocation of time during the diving cycle. Behav. Ecol. 3, 255–265 (1992).Article 

Google Scholar 
79.Thompson, D. & Fedak, M. A. How long should a dive last? A simple model of foraging decisions by breath-hold divers in a patchy environment. Anim. Behav. 61, 287–296 (2001).Article 

Google Scholar 
80.Aoki, K., Sato, K., Isojunno, S., Narazaki, T. & Miller, P. J. O. High diving metabolic rate indicated by high-speed transit to depth in negatively buoyant long-finned pilot whales. J. Exp. Biol. 220, 3802–3811 (2017).PubMed 
Article 

Google Scholar 
81.New, L. F. et al. The modelling and assessment of whale-watching impacts. Ocean Coast. Manag. 115, 10–16 (2015).Article 

Google Scholar 
82.Parsons, M. J. G., Duncan, A. J., Parsons, S. K. & Erbe, C. Reducing vessel noise: An example of a solar-electric passenger ferry. J. Acoust. Soc. Am. 147, 3575–3583 (2020).ADS 
PubMed 
Article 

Google Scholar 
83.Dale, S. J., Hebner, R. E. & Sulligoi, G. Electric ship technologies. Proc. IEEE 103, 2225–2228 (2015).Article 

Google Scholar 
84.Anwar, S., Zia, M. Y. I., Rashid, M., De Rubens, G. Z. & Enevoldsen, P. Towards ferry electrification in the maritime sector. Energies 13, 1–22 (2020).Article 
CAS 

Google Scholar 
85.Filby, N. E., Christiansen, F., Scarpaci, C. & Stockin, K. A. Effects of swim-with-dolphin tourism on the behaviour of a threatened species, the Burrunan dolphin Tursiops Australis. Endanger. Species Res. 32, 479–490 (2017).Article 

Google Scholar 
86.Sprogis, K. R., Bejder, L., Hanf, D. & Christiansen, F. Behavioural responses of migrating humpback whales to swim-with-whale activities in the Ningaloo Marine Park, Western Australia. J. Exp. Mar. Bio. Ecol. 522, 151254 (2020).Article 

Google Scholar  More

Results agreed closely with Ref.32 when wind speed-specific drag coefficients were introduced that seemed to be reasonable, as the latter lie slightly under the values published by Ref.17 for a stiffer Canary date palm (Phoenix canariensis). Notwithstanding, the model’s outcome on its own would not be a valid contribution to the field of palm biomechanics and risk assessment. Hence, and from here onwards, the following observations are deemed to be crucial:The researcher32 used data on the mechanical properties of green tissue of senile 80–100-year-old coconut palm stems harvested in 2010 in Fiji and Samoa and which were then published in Ref.34. His model predicted that the critical wind speed for failure of the stem fibres was 82.8 km/h (23 m/s). However, for instance, the cyclone “Val” hit Samoa with wind speeds up to 140 knots (259.28 km/h) in 199135. The real palms from Ref.34 that served for Ref.32 must have withstood, along their life span of 80–100 years, many times wind speeds that exceeded the theoretical “critical wind speed” of 82.8 km/h as predicted by Ref.32. And yet, those palms still stood upright when they were harvested. And no mechanically-damaged tissue in the harvested cocostems was reported in Refs.[32, 34], which suggests that those coconut palms had not suffered any failure of their stem fibres, not even when wind speeds up to 259.28 km/h hit the island they were growing on. This observation adds up to others (e.g.10,19,20,33) that suggest that the engineering approach as used by Ref.32 and used, albeit much simpler, in e.g. the “tree-statics” of Ref.3 may have a limited predictive value.Next: Lowest FI was calculated here at a height of 13.1 m up the stem, whereas32 predicted failure between 6 and 10 m at 23 m/s (while the failure area would move towards the stem base with higher wind speeds). The reason behind this disagreement is simple: the herein employed model uses simple beam theory, which assumes the deformation or curvature of the stem is null. And32, on the other hand, depicted the bending over of the slender palm with a highly-pronounced curvature in the lower half of the stem, which would, hence, cause greater stresses. And as this happens, the upper half of the stem would align itself more with the wind and thus be less deformed, while also the experienced stresses would be lower there (see the figure in Ref.32 p. 126).Furthermore, the present model “fabricated” the rising stresses, by modeling the stem as a hollow wind turbine tower with a changing t/R related to wind speed. The rationale behind this approach is the following: the beam theory neglects that, in palm stems, stresses increase exponentially from the neutral axis and that non-linear deformations and strong curvatures can be experienced by slender palms in high winds, as reported by e.g.23,32. This is thus one of the arguments against the simple beam theory in slender trees and palms, and which has been used in the present model and by e.g.1,3,4,5,16,17,18. One of the premises of that theory is that deformation of the beam should be small. And if, for instance, a slender Mexican Fan Palm bends over, the curvature of the stem could be too pronounced to be faithfully modeled with that theory. And real stresses due to that curvature would hence be higher than predicted. For instance, a company markets commercial pulling test on palms and shows it application on a 22-m-tall and leaning Mexican Fan Palms (La Aduana, Málaga)36. And theirs may thus be an untenable and risky suggestion. A crucial observation can be made here: if regular beam theory (as used in commercial pulling tests and wind load analysis software packages) had been blindly used for the coconut palm simulation (i.e. assuming a stiff and solid beam with no sharp rise in stresses due to the pronounced curvature of the stem and the bending stress increasing linearly from the neutral axis as if the cross-section were both full and isotropic), breaking safety factors would have been greatly overestimated (nearly doubled at 60 m/s) as stresses would have been underestimated. Which could lead to dangerous, and deadly, situations in real-life palm risk assessments.Next: As mentioned in Ref.20 the results of Ref.32 were not validated experimentally. Hence, and even barring structural defects such as cracks or decay, it is a theoretical model that may not allow for accurate safety predictions yet. Furthermore, palm stems were reported to break either at mid-stem or just below the crown37. Whereas32 predicted failure well below mid-stem, extending towards the bottom of the stem at higher wind speeds and so, his results do not agree with real failures as reported by Ref.37. Also, the Red Palm Weevil (Rhynchophorus ferrugineus) has been said to affect the structural stability of palm trunks by excavating tunnels which could lead to their collapse38. The loss of structural integrity due to this tunneling is a type of defect that is not taken into account in Ref.32 either. And Ref.39 suggested several factors that may influence breakage of coconut palms subject to cyclones, and not all are taken into account in the simulation: e.g. the ratio between diameter of the bole and stem height, different mechanical failure modes related to certain hybrids (e.g. the of Malayan Red Dwarf versus Tall Palms), biomechanical degradation due to Phytophthora palmivora and crown characteristics (weight and volume of fronds and crop). The latter also reported fracture of the bole at the root-soil plate level and below ground, while other palms were left leaning after partial uprooting and others had their stems broken at different heights39. This suggests that the predictive value of the approach used by Ref.32 is rather limited and that, even though his is an impressive and highly-valuable contribution, it may not be extrapolable to real-life palm risk assessments yet.Next, the commercial pulling tests and wind load analysis software of Refs.4,5 have been used on palms in Spain. Nevertheless, no clear references pointing towards scientific, peer-reviewed papers of theirs could be found in those publications, relating to scientifically sound data and scientifically contrasted procedures that would support their methods. Which, surprisingly, stands in sharp contrast with their criticisms toward a competitor in business: that the latter would offer (similar) methods without adducing supporting scientific evidence4,5. Also an influential publication asserted with a “generalised tipping curve” that the critical threshold of tilt angle for uprooting would be 2.5° (assertedly based on 400 trees), which seems to be the foundation of their commercial pulling tests3. That strong claim was later questioned as it still seemed unproven and thus hypothetical19. Moreover, the asserted results of3 stand in sharp contrast with those of recent researchers who, after two million measurements on more than 8000 trees, do not assert to have found any critical threshold of tilt angle yet40. Pulling tests have been seriously questioned of late and it is not clear to what extent their predictive value in dicotyledonous trees is reliable or not19,20. Which makes their use for monocotyledonous palms even more questionable as the latter are biomechanically very different from trees. Pulling tests were developed from pulling over dicotyledonous trees, with their corresponding characteristics (regarding material and geometry) of their root system and stem base. However, and on the other hand, the fleshy palm roots tolerate significant bending and twisting before they undergo mechanical failure41. And the roots also sprout from the stembase in a way that is similar to onions. Moreover, the effects of organic exudates from the roots on the aggregation of soil particles, which would thereby create a cement-like soil consistency, was reported too41. And these exudates have seemingly not been reported for dicots yet. And, thus, recommending pulling tests for palms seems to stand for blindly extrapolating hypothetical (as seemingly yet unproven) values for dicotyledons, into the unknown and unexplored realm of palms, and this is could thus be deemed very questionable.Next, destructive experiments were experimentally carried out with the pulling test method of Refs.3,45 on a hollow date palm (Paseo Marítimo, Mataró, Spain) that presented a thin residual wall and a large longitudinal opening42. The palm was pulled with a fixed maximum load of 1.5 kN while performing 20 consecutive measurements with a strain gauge sensor or “elastometer”. The measuring tools (elastometer, inclinometer) had been provided by Refs.3,45. The distance between the two pins of the strain gauge sensor was 2000 mm. The direction of the pull was perpendicular to the opening of the cavity. The first 18 measurements were carried out by placing the sensor aligned with the stem (longitudinally) and in the direction of the pull, conform to the classic pulling test procedure. The highest longitudinal deformation (aligned with the stem) recorded was 0.089 mm (at 2 m height) and 0.045 mm at the height of the cavity (1 m) with a static pulling load of 1.5 kN and conform to the classic pulling test procedure of Ref.3. However, two alternative measurements were made afterwards by placing the sensor first in an angle of 90° (horizontally and bridging the cavity opening) and then in an angle of approximately 30° over the open cavity (Fig. 1), to assess if there could be any shear. Then, the strain gauge sensor recorded the astonishing value of 0.321 mm at the same pulling load (1.5 kN) when placed obliquely (30°) over the open cavity. Which is more than seven-fold the maximum axial strain measured at the same height of the stem and aligned with the pull, which indicates that shear (and not longitudinal strain) was the highest. The two sides of the open cavity seemed to slide over each other. This phenomenon can be visualised by bending a softcover book: the pages slide over eachother. Palm wood is much weaker regarding shear stresses and when this is coupled with extraordinary shear deformation (due e.g. a large, open cavity such as here) failure of the hollow stem can be triggered at lower loads than predicted by the beam theory. And if this extremely hollow stem exhibited high shear deformations under a transverse pull, then logic says that also other structural behaviours (e.g. ovalization, cracking or kinking) could set in.Figure 1The strain gauge sensor was placed obliquely over an open cavity in a date palm, recording high shear deformations. The red arrows show the direction of the strain, producing shear as the two halves of the stem seemed to slide over each other, while a static pull was being applied.Full size imageTo continue: the death of a man crushed by a Canary date palm in 2020 in Barcelona (Spain), triggered the implementation of drilling with the Resistograph of Ref.1 and “oscillation tests” on 2026 date palms with a “reliability of almost a 100%”, carried out by Refs.36,43. Nevertheless, it was stated that drilling cannot predict the residual strength of a structurally damaged trunk44,45. And it was shown that boring into decayed zones would probably augment the speed at which decay spreads46. Micro-drilling would allow fungi to grow out radially due to the microenvironment the narrow channel creates47. And drilling was also questioned by Ref.48. And correct assessments would rely on comparing results with both known standards and decay-free cores taken from the same tree, which would make this method thus highly-invasive49. Hence, (micro)drilling is currently highly questionable. And it was reported that the “oscillation tests” of that company were as follows: the palm is pulled manually with a rope and if the whole stem moves the palm stem is regarded as sound50. This was the only description found of those “oscillation tests” on palms after exhaustive online literature review. That company recommends that their oscillation test be used in all palm risk assessments36. Nevertheless, no clear references can be found in Refs.36,43 that would point to peer-reviewed publications in scientific journals that would validate their claim, so it is not clear if it is scientifically tenable or not. However, their claim could be interesting, if robust and supporting scientific evidence could be adduced and if their procedure were unbiased, documentable and reproducible by a third party.Next, the famous “70%” criterion was adduced by the Municipal Manager of the City Council of Barcelona after the deadly accident with a Canary date palm and the manager was reported to have said that inner decay was considered problematic if at least 70% of the stem diameter was affected by it12. And that the risk of breaking would gain importance if the extent of decay occupied at least 70% of the stem radius13. However, if the palm collapsed with only 25% of its radius affected by decay, should one then not immediately refute the “70%” criterion? Notwithstanding, the applicability of this criterion as suggested for palms by Ref.1, can easily be refuted scientifically too: Firstly, suspicions of falsification were published regarding that highly-influential Visual Tree Assessment (VTA) rule t/R = 0.32 or 70%51. That t/R rule for risk assessments defined the (supposedly) allowed degree of hollowness of a tree trunk and is still being used world-wide, although it is allegedly the result of falsification51. Secondly, that famous “70% criterion” was, moreover, developed for dicotyledonous trees and not for monocotyledonous palms. Thirdly, tangential MOR (tension perpendicular to grain) in coconut wood was found to be as low as 0.233 kN/cm2 whereas longitudinal MOR could reach 5.22 kN/cm252. And shear strength of date palm wood could be as low as 7.14% compared with its longitudinal MOR52. Even longitudinal MOR varied greatly between coconut, oil and date palm52. For oil palm (Elaeis guineensis), the proportion of tensile strength perpendicular to grain to longitudinal MOR was reported to be only 2.08%53. And another researcher found that tangential MOR would be 36.77% approximately of longitudinal MOR in coconut palms32. Therefore, tangential cracking followed by longitudinal splitting in straight hollow stems would thus be triggered earlier in date and oil palms than in coconut palms. Palm wood is highly anisotropic compared to dicotyledonous hardwood trees and, above all, there are great differences both among different palm species and even within the same palm species, depending on e.g. age and growth. And, hence, a fixed t/R rule seems to be an untenable recommendation for palms. And this would even not be applicable in stems with side openings or those with only heart decay30. Or with irregularly-distributed pockets of decay combined with cracks or invaginations. Even if one dared to leave aside the fact that high stresses can be caused by strong curvatures of slender palms in high winds (see e.g.23,32) and those are not taken into account in the VTA t/R rule of 70%. Cross-sectional flattening or ovalization, leading to cracking of thin-walled hollow stems, is neglected by classic beam theory too, while those structural failures depend highly on the MOE of the wood54. And MOE, MOR and density vary greatly according to species, age, et cetera and thus preclude a fixed t/R rule entirely from being useful when the aim is to predict the structural collapse of (especially) palms.Next: Scientists published breaking safety factors and critical wind speeds for severely decayed date palms in a highly-frequented area in a major Spanish city16. Their suggestion that those decayed palms would withstand wind speeds of at least 135 km/h even made headlines55. They did not fully explain the underlying methodology in their online-published report. However, the manual of the acoustic tomograph (Fakopp56) they used, suggests that breaking safety factors and critical wind speeds were calculated by means of simple beam theory, which is based on longitudinal stress and strength, and assumes that wood is a homogeneous material (i.e. isotropy). However, palm wood is highly anisotropic, meaning that shear, delamination, torsional, radial and tangential tensile stresses could cause structural collapse, especially in decayed palm stems, even if the beam theory were applicable. So, their claim that decayed palm stems in a market square would withstand hurricane-type winds, is possibly based on what looks like a methodological error on their behalf. On the other hand, a method for the breaking risk assessment of Canary date palms was offered in a scientific paper through beam theory and a hypothetical static wind load17. But, fortunately, the latter acknowledged that their approach could not be valid for decayed stem areas due to shear and that failure due to progressive fatigue was not considered either. And a highly-cited researcher calculated longitudinal stress in a Mexican fan palm (Washingtonia robusta), seemingly based on the assumption that bending stress would increase linearly from the neutral axis and by means of a static pulling test combined with simple beam theory18. However, the herein offered remarks suggests that the latter’s approach for palms may be a simplification that has room for improvement too. A Spanish researcher published results from static bending tests of planks, sawn from a Canary date palm, to be used in the context of mechanistic risk assessment models57. Unfortunately, the procedure he used precludes his results from being useful in that context, as current models need MOR and MOE values obtained from compressive axial and tangential tensile tests. Nevertheless, in his review he rightly acknowledged the dubious efficiency of published risk assessment models, as they would depend too much on a wide palette of unknown variables (e.g. Cd, MOE, MOR, density) and that decay detecting tools were not efficient, as reference values did not exist (e.g. to calculate strength loss in comparison with sound palm wood)57.Next: Some readers will surely feel tempted to use the model for, and extrapolate the results to, commercial palm risk assessments. But that would clearly be premature, as can be inferred from the following observations: a wind-speed-specific drag coefficient was proposed herein to simulate the coconut palm of Ref.32. But a sturdy stem (in contrast with the studied 25-m-tall and flexible stem) would need to be modeled with a different Cd. And to make Cd transferrable to other palms, non-linear deformation would have to be taken into account. This non-linearity is the result of the slenderness, anisotropy and geometry of the stem, the flexibility of the crown and overall out-of-phase damping. A greater curvature of the stem leads to higher three-dimensional stresses (longitudinal, radial and tangential). And Poisson’s ratios also determine deformation of the fibres32 and this may differ too. And all of this clearly exceeds the capacities and predictive power of simple beam theory. Also, wind drag has commonly been estimated as being proportional to the square of the wind speed. However, it was shown that this estimation may be too high for flexible culms, as at higher wind speeds the drag would be linearly proportional58. And, hence, real loads would be lower than predicted. They also found that the risk of mechanical damage was comparatively lower at higher wind speeds, as the plants’ height was reduced by up to 45%58. It was suggested that coconut palms would resist hurricanes better than dicotyledonous trees because of the same strategy32. On the other hand, common sense and observation suggest that the mass of palms (stem, crown and crop) combined with violent gusts may lead to dynamic loading that far exceeds predictions that take into account static loading only. Two simulation studies did not consider dynamic loading, damping, looping or inertia effects20,32. Nevertheless, the results the first agreed very well with commercial softwares that, assertedly, would include dynamics and natural frequencies20. For instance, it was asserted that “and statics integrated methods that combine static pulling with dynamic wind load assessment (Wessolly 1991; Brudi and van Wassenaer 2002; Detter and Rust 2013)” (sic)59. Related authors also suggested that a natural frequency factor was incorporated in their calculation of the wind load and bending/uprooting moment of their pulling tests3. However, robust scientific evidence that would support their claim was not found and neither did the mathematical simulations find any evidence of dynamics20. Not including the influence of dynamics (e.g. the swinging of slender trees and palms) in a wind load analysis could underestimate real wind-induced loads. Which means that the palm or tree could thus fall down even if it had been assessed as “safe”. Also the weight of crop (e.g. dates or coconuts) could add inertial forces to the swinging and this could be a subject for future research on wind loads in palms20.Further: Mechanical properties (strength, stiffness and density) of green palm tissue are still a relatively unexplored field, although several palm species have been studied17,22,32,52,53,60,61,62. Properties from these publications of other palm species could be introduced in the model to explore their importance relative to other influencing factors such as slenderness, wind speeds, loads, et cetera. But, and even though this procedure has led to good agreement for FI of Ref.32, more research on the applicability of the model should be carried out.Also: The herein used approach is based on a simplified version of the theory of elasticity, which ignores stress concentrations (e.g. around knurls or defects in wood), Inglis’ potential energies, fatigue and crack propagations as described by Ref.63, which can lead to unexpected structural collapse if one relies solely on simple beam theory. The need to explore those ideas was suggested, as understanding their influence could be the key to understanding the relationship between structural failure and wind42. Fortunately, those relatively unexplored ideas were later applied to calculate critical wind speeds for failure in forest trees64. This could thus be an interesting starting point for research on the breaking prediction of palms and trees.Moreover, the beam theory as used by some of the herein mentioned authors is aimed only at predicting conventional bending failure (axial compression stress that exceeds MOR), while low t/R ratios can lead to Brazier buckling or tangential cracking followed by longitudinal splitting in hollow stems30. The formulations were offered to predict the bending moment at which cracking failure would occur in hollow trees, based on t/R, MOE and tangential tensile MOR54. So, and for instance, if one took an oil palm and a coconut palm, both hollow and with an identical t/R and wind loading, the first would crack earlier than the second due to a lower tangential MOR. And as t/R decreased, failure modes would be bending failure, cracking and Brazier buckling respectively, for oil palm. Whereas in coconut palm, and depending on t/R, bending failure would occur earlier than the other two modes due to a comparatively higher tangential MOR/longitudinal MOR proportion. Which is also evidence why fixed t/R rules (e.g. 0.32 or 70% of the radius) and beam theory (e.g. pulling tests) cannot be applicable to palms. Hence, incorporating cracking and buckling predictions in the assessment of decayed and concentrically hollow palm stems, could also be an interesting lead.Next, it was found that the Brazier calculations (based on MOE) agreed with the BS outputs (based on MOR) of the model for all heights along the stem and all wind speeds, when the stem was modeled as untapered, which is interesting. The relationships of varying MOR and MOE along the stem were based on the measured densities of Ref.32, so there seems to be a consistent mathematical relationship between the MOR, MOE and density values32. At first sight, densities of green palm wood could thus be an interesting future research subject. However, and on the other hand, it was reported that no correlation existed between density and mechanical properties in date palm wood52. Which would make future mechanistic modeling thus even more challenging.Next: None of the herein investigated models and criteria fulfill the requirements (i.e. that models should account for cell wall expansion and sclerification as a result of height growth and age) stated by Ref.60. And they could therefore be precluded from being useful for palms, as the latter both grow and age.Researchers also recorded longitudinal tensile stress on the surface of upright growing trunks, whereas compression stress was found at the bent area of leaning trunks in coconut palms due to growth stresses65. They also found compression stress in the outermost portion of the inner cylinder of the coconut stems, which they said was radically different from dicotyledonous and coniferous trees. So, this also questions the application of fixed t/R rules, pulling tests or wind load analysis combined with beam theory on palms, as those methods neglect growth stresses and their biomechanical importance. For instance, if the central cylinder were missing (e.g. due to butt rot caused by Ganoderma zonatum) then the lack of those inner and outer growth stresses and strains should be accounted for.However, now we will suppose, and for the sake of the argument, that we approached the pitfall in which some of the aforementioned companies and researchers have seemingly already fallen. At a first glance, it would be appealing to suggest the following method: consider that commercial software packages for wind load and breakage predictions were successfully simulated20. And that special software packages were also suggested to accurately measure the vertical area of e.g. a palm crown, which would thus allow to perform a wind load estimation that would meet the standards of the commercial software packages investigated20. Suppose a wind speed-specific drag factor be introduced, such as proposed for Canary date palms by Ref.17 or the one found here for coconut palm. And that the formulations for the critical bending moments for tangential cracking from Ref.54 and the ones employed in this study for pure bending failure be incorporated, together with the wood properties as published for several palm species by e.g.17,22,32,52,53,60,61,62. Furthermore, values for peripheral material properties were obtained from the ring that corresponds to the outer third of the radius32,52. And take a non-linear bending stress distribution in the cross-section of the stem, which rises exponentially from the neutral fibre to the peripheral outer ring made of the most dense, stiffest and strongest tissues21. Then, a simplified assumption would be to calculate stresses taking into account only the outer third of the radius (i.e. t/R = 0.33), as if it were a hollow wind turbine tower and as has been done in the present study. And this, to simulate (in an extremely simplified manner) non-linear bending stress and peripheral material properties (note: it should absolutely be stressed here that this is not regarded as a validation of the VTA t/R = 0.32 rule, as the rationale for its use in the model differ from the rationale of Refs.1,15, while the inapplicability of the latter’s claim has been amply evidenced in this study). In this way, theoretical safety factors could then be calculated and compared for bending versus cracking failures of the hollow palm stem, for varying wind speeds and several palm species. This would thus be similar to the widely-cited Statics Integrated Assessment (SIA) and Statics Integrated Methods (SIM) of Refs.3,45, but then for palms and slightly enhanced (as it adopts cracking failure, the varying material properties across and along the stem and a wind speed-specific drag factor). And as less advanced methods (e.g.1,3,16 have already been commercially marketed, a non-scholar could perhaps be tempted to commercialise this model in a software package or use it for their consultancy services too. However, this approach would still suffer from the same limitations as described in Refs.19,20 and in the present study. And it would still be theoretical, as the variables concerning the structural stability of hollow trees and palms may be too diverse to be assessed with current methods19. And the combination of small deviations in the real palms from e.g. the published values for MOR and MOE and theoretical drag factors (and hence predicted wind loads) could result in a global deviation that may invalidate the outcomes (the latter concerns all of the herein investigated methods too)33. Hence, and even though it is not the corresponding author’s idea to wholly negate the usefulness of the herein investigated methods, it is crucial to point out that both their validation and predictive value are seemingly problematic.A crucial rationale for presenting the utterly simplistic model in this paper was the following: supposedly complex models such as e.g.3,4,5,45 or the advanced 3DFE simulation of Ref.32 may obscure that fact that those models can be as tied to the same limitations as the simple model presented herein. And apparently complex equations (or e.g. a high number of citations of the related papers) may deviate the readers from the fact that factual empirical and scientific evidence could still be missing that would validate the models for real-life purposes. Hence, a simple model such as the herein presented one, may serve the purpose of pointing out the flaws and limitations of the seemingly more advanced models, while it even seems capable of simulating internationally-renown commercial software and methods20.So, the time seems to be ripe now to go beyond the classical procedures as trusted upon by the arboricultural community so far (and discussed before).Even in straight, thin and idealised cantilever beams, bending–torsion coupling deformations can arise due to the dissimilar bending stiffnesses when the two planes (horizontal/vertical) of the cross-section are of uneven dimensions (instead of a e.g. a perfectly circular or annular cross-section)66. Palm and tree stems are not always perfectly round due to dissimilar diameters in the horizontal versus the vertical plane (e.g. in cases of reaction wood, open cavities or uneven radial growth due to touching physical obstacles). Hence, simple beam theory (e.g. pulling tests) may thus not account for torsional (and, ultimately, catastrophic) behaviours, even in straight stems. Moreover, if improperly applied, simple beam theory may theoretically predict the strength and stiffness requirements of a structure to be satisfying, while unforeseen collapse may later occur because of the loss of stability (buckling), including intriguing phenomena such as non-linear geometric deformation and wrinkles66. Translated into arboricultural language: the tree or palm that had been assessed as “safe”, suddenly collapses unexpectedly. Therefore the need in this paper to show the arboricultural community that structural collapses, that have been studied for centuries in other fields such as mechanics and engineering, should not be ignored.The risk of buckling of a Mexican Fan Palm (Washingtonia robusta), assessed by the corresponding author in 2003 in the Atocha train station (Madrid, Spain), gave birth to a proposal to assess the risk of Euler buckling while carrying out wind load estimations in order to optimise artificial supports (e.g. cabling of the palm to nearby structures)28. Prior to the assessment of the last standing palm, several other slender Mexican fan palms in that train station had already collapsed, even though the interior of this giant greenhouse is free of wind loads (Fig. 2). The photograph is a testimony to a rather neglected fact in commercial arboricultural methods: structural collapse in absence of wind loading and pure post-buckling failure. In this case it was hypothesized that these palms had initially become elastically unstable, by exceeding their critical stem height and weight. It was hypothesized, too, that this had been caused by their unlimited growth towards the glass ceiling searching for light, the absence of external loading stimuli such as wind (the lack of which would have made the palms not to invest in stiffer and denser wood) and optimum growing conditions (permanent moisture and warmth). The weight of the crown, small horizontal displacements, watering from the ceiling (i.e. fog to keep the atmosphere moist) and resulting gravity forces would then have further influenced the failure process, leading to final collapse. This example illustrates how plants can adapt to their environment and that biomechanical failure can be possible in total absence of wind loading. In large-wave Euler buckling, the column curves and deviates laterally to escape from compressive loads (such as e.g. self-weight) before axial stresses surpass axial MOR. The column becomes elastically unstable and buckles under its own weight. The critical weight divided by self-weight gives the safety factor and only when this safety factor is higher than unity can columns, or plants, bear additional loads such as wind, snow or ice. The critical buckling height or weight is a function of stem height, diameter, tapering, MOE, density of the wood and loading conditions27. The latter also showed that buckling safety can be overestimated if the stem is improperly assumed to be untapered, cylindrical, free of imperfections and isotropic27. They also offered an overview of why predictions of structural collapse may easily differ from real-life situations27. Moreover, the bifurcation point is the sudden jumping process of a beam from a straight-line to a bent shape, causing instability or buckling66. Pre-buckling analysis has proven to be rather straightforward for a simple pole, while the post-buckling process that describes the finite deformation of the structure (which may lead to its collapse after damage and faults accumulate to a certain value) requires a large set of numerical solutions67. Strong geometric nonlinearities and large displacements of the post-buckling behavior of a slender rod were studied, leading to a quantitative calculation of the post-buckling deflections of a hollow oil sucker rod67. Translated into the world of palm biomechanics, it means that: while pre-buckling of the stem would already be a daunting task due to the varying taper and MOE, predicting its post-buckling behaviour and final collapse (including structural faults such as e.g. cracks or pockets of rot) seems to be out of reach, as palm stems are not human-made structures. And yet, it seems reasonable not to ignore this type of structural behaviour in future palm risk assessments. It was acknowledged too that Brazier buckling played a crucial role in the local instability of plant stems66. And this was also a reason to include Brazier buckling of a hollow wind turbine tower to simulate breaking safeties of the coconut stem in the present paper.Figure 2The slender Mexican fan palm anchored to the ceiling of the Atocha train station, Madrid. The cabling configuration was installed to minimise damage in case of post-buckling collapse. The other palms had collapsed before, even though there are no events of wind inside this giant greenhouse.Full size imageNow, and as a second part of this “Discussion” section, the following reasonings elucidated from literature overview, visual observation and intellectual reasoning, are presented to postulate ideas that may serve to show the way towards a future unified theory on palm risk assessment.First: The biomechanical structure of palms seems to have evolved towards highly-efficient energy dissipation and viscoelastic damping capacities under strong and dynamic wind loading. To achieve this, a triple-helical mesh of tough (high tensile strength) fibrovascular bundles is embedded in a soft parenchymatous foam, which both contribute to damping and energy dissipation32,68. The fibrovascular bundles run along the stem in a screw-like fashion and across the stem in a radial zigzag pattern (this also sets palm wood also apart from dicotyledonous wood, as in the latter the fibres are stiffly glued together and, most importantly, axially aligned). It was asserted that this screw-like pattern can hold the bending stem together under high wind loading as it lends the stem a higher stiffness and strength when the fibrovascular bundle orientations varied between 0° and 9°32. This pattern was also suggested to minimise longitudinal splitting and thus enhance the mechanical efficiency of the stem32. This structure was an inspiration for spirally-laminated hollow veneer-based composite poles32. Also high microfibril angles across the fibre cap would result in a high extensibility of the stiffening tissue, which would enable palms to cope with considerable deformations under wind loads in Mexican fan palms69. Large deformations in bending and torsion under wind loads of the petioles were said to combine efficiently with water and nutrition conduction, due to the optimized connection of their vascular bundles to the leaf traces68, which allows to suppose that also the crown is optimized regarding damping and energy dissipation brought about by dynamic winds. And the contribution of both parenchymatous and vascular tissue of palms to energy dissipation, dynamic response and flexibility, and thereby improving impact resistance, was described too70.Second: Palm wood is highly sensitive to shear, delamination and splitting in comparison with dicotyledons. For instance, when samples were taken by Ref.17 to perform longitudinal compression and tension tests, then this irregular structure of the palm tissues unwillingly led to longitudinal fractures, sliding and shear in the samples, and thus seriously limiting experimental data on axial MOR. And thick disks of coconut wood were manually torn apart, while the delamination followed the helical pattern of the fibrovascular bundles that tangentially deviated across the disc diameter32. Hence, it is thus not unreasonable to suppose that this sensitivity to delamination and shear may lead to the stem’s structural collapse, especially when this helical path of bundles is interrupted by a mechanical defect (e.g. pockets of rot, irregular decay, cracks or tunneling by Red Palm Weevil). This reasoning aligns with another researcher’s too, who likened coconut and oil palm stems to a composite material made of a matrix and reinforced elements and found that shear and tension perpendicular to grain greatly govern the bending behaviour and structural stability of the stem52. The aforementioned “spirally-laminated hollow veneer-based composite poles” suggested by Ref.32 may be very stiff and strong when undamaged (i.e. if this helical pathway of fibrovascular bundles is not interrupted by a mechanical defect and thus a completely defect-free beam). But, an interruption along this path may trigger delamination and splitting along the “veneer”. Crack propagation and splitting could thus follow the helical path of the fibrovascular bundles. And predictions based solely on simple beam theory and axial stress and strain would then be less than acceptably reliable. Observations and experiments that seem to support this hypothesis are e.g. the aforementioned Canary date palm that crushed a man in Barcelona, as a small inner crack was said to have triggered the sturdy stem’s collapse with a breeze of only 38.2 km/h10. Also pulling test experiments carried out in 2004 by the corresponding author showed that the mechanically damaged palm stems under artificial loading started splitting first, leading to full collapse afterwards42. Those experiments (partially published in 200542) had been kindly supported by Josep Selga S.L., the City Councils of Terrassa and Mataró and the Asociación Española de Arboricultura, while the instrumentation had been kindly provided too (Picus tomograph: L. Göcke Argus Electronics; Pulling tests: Brudi and Partner Tree consult and Dr. Ing. L. Wessolly; Resistograph F300, IML: the City Council Terrassa). The aim was to assess whether the pulling tests of Refs.3,45 could be adapted to palms or not and if experimental data for MOE could be obtained from standing palms. Acoustic tomography (Picus tomograph) and microdrilling (Resistograph F300) had also been carried out on several damaged palms, but had not facilitated any reliable breakage prediction either (unpublished results). An example is shown in Fig. 3 where a desert fan palm (Washingtonia filifera) collapsed under a static pull, after slanted longitudinal splitting and delamination was initiated at the border of the open cavity (upwards and downwards)42. Also Fig. 4 shows how delamination (triggered at the height of the open cavity under a static pull) led to total collapse of a date palm stem. No primary axial compression failure was observed macroscopically42. And a hollow date palm exhibited extremely high shear values in comparison with axial deformation at the height of a large, open cavity (Fig. 1)42. Moreover, it is not unreasonable to suppose that if strong, cyclic and repetitive dynamic wind loading had beaten these three palms (instead of a static pull), the risk of structural failure could have been heightened by progressive fatigue of the wood around the structural defects (and thus earlier crack formations/propagations and at lower loads than with the static pull).Figure 3When a decayed desert fan palm stem was statically pulled, collapse was initiated by splitting of the hollow stem. Cracks first appeared above and below the open cavity and initiated at its borders (red arrows) and total collapse only ensued after large longitudinal splitting and delamination.Full size imageFigure 4When a decayed date palm stem was statically pulled, splitting was initiated at the open cavity and total collapse ensued by delamination.Full size imageThird: Highly deformable and soft but elastic materials can exhibit types of structural deformation under mechanical loads that are unlike those commonly observed in elastic structures that behave linearly71. Kinking at the inner side of soft, elastic cylinders was observed after the cylinders had become elastically unstable due to Euler buckling71. The extreme localization of curvature at the compressed inner (not outer) side exceeded a critical value leading to a sharp fold. When the cylinder was kept under a bending load for several minutes, irreversible defects appeared at the location of the inner kink which, in subsequent loading cycles, progressively lowered the cylinder’s structural stability under the same amount of load71. Translated into palm stems, and assuming they are highly deformable, soft and elastic, this means that inner kinks and defects could appear and lead to structural collapse due to fatigue and cyclic loading beyond the critical curvature. Brazier buckling was also observed in soft, elastic and hollow cylinders and the occurrence of either kinking and/or ovalization was found to be dependent on the ratio between the diameter and the wall thickness71. When one envisages palm stems as has been done in the present paper (a viscoelastic cylinder), then the kinking and ovalization of the cylinder (here: the palm stem), after becoming elastically unstable, could thus lead to abrupt structural collapse while not obeying simple beam theory. Calculation of the critical curvature at which buckling sets in was said to be rather straightforward, but the posterior evolution of the kink or defect would need detailed non-linear theory71. Hence, the modeling of elastic pre-buckling (i.e. prior to these aforementioned structural failures) seems to be more within reach for palm stems than post-buckling collapses. No experiments have been performed on palms yet to either confirm or refute these extrapolated suggestions, but the latter are possibly worth considering in future research or risk assessments.Fourth: Developing a mechanical model seems currently out of reach as strength and stiffness (and thus damping) seem to evolve over time in the palm stem as a function of the location of the vascular bundles within the trunk, age (and ensueing additional cell wall layers and (secondary) growth within the trunk) and growth conditions52. Also the lack of a statistical correlation between MOR and MOE and wood density in date palms is, inexplicably, contrary to other investigated palm species, which also obstructs the path towards reliable mechanistic modelization52.Fifth: It was stated that “Reliable prediction of delamination growth is still proving to be problematic” in human-made wood products, whereas simply localising starting points for delamination would possibly be more within reach72. From which it can thus be inferred, that reliable predictions of delamination-triggered collapses of Nature-made palm and tree trunks seem currently to be out of reach. But that would still be no reason to neglect this type of structural failure).Sixth: The existence of silica in palms was mentioned by Ref.52 (p. 158) and studied by e.g.73,74. Researchers concluded from a literature review that the mechanical properties of palms could be enchanced by silica73. And the role of silica in plants was described as: “Biomineralization is a naturally occurring process by which living organisms form skeletons from inorganic minerals such as silica and calcium”75. The latter also found flexural rigidity in rice plant leaves to increase with increasing silica content. It has been suggested by practitioners and arborists in Spain that silica and biomineralization would make the palm stem stiffer and stronger around structurally defective areas, as an alleged reaction to strength loss percieved by the palm itself (i.e. a substitute for compensation or thigmomorphogenesis as studied in dicotyledons), but no scientific findings were found that would support their suggestion.Seventh: Local mechanical performance (i.e. damping and the diminution of stress discontinuities) of a Mexican fan palm stem could be controlled by the plant itself up to a certain point by adaptation69. Which would further complicate the mechanical modeling of structural stability versus (wind) loads.Eighth: The cracking formulation of Ref.54 should unfortunately be precluded from being useful in hollow palms, as their formulation assumes that the fibres are aligned along the tree axis, while palms present a mesh of triple-helical fibrovascular bundles in a screw-like pattern along and across the palm stem.Ninth: Based on visual observation, young and still flexible and soft Mexican fan and windmill (Trachycarpus fortunei) palm stems seem to exhibit a viscoelastic behaviour when manually pushed and pulled. Their moving out-of-phase with the pulls can be felt by hand and feels like a structure made of foam, but with a certain resilience. Their behaviour resembles neither that of a steel spring nor that of foam or a stiff and non-deformable beam. And this in contrast with e.g. flexible dicotyledonous saplings and tree branches that almost behave like springs or lashes when laterally loaded and released by hand. Also visual observation of the damped manner in which older, taller and stiffer Mexican fan, date and windmill (Trachycarpus fortunei) palm stems move out-of-phase in strong winds seems to confirm this. And in several palm species the woven mesh of leaf sheath palm fibres attached to the stem also exhibits a damping and viscoelastic behaviour when manually manipulated. In windmill palm for instance, the stem is wrapped in a burlap-like mesh of brown and coarse leaf sheath fibre, clasped around the trunk. Manual manipulation of that mesh suggests that friction among the fibres could contribute to damping of leaf and stem movements. A review of published findings on damping and energy dissipation in palms seems to confirm these visual obervations too (see32,68,69,70). A viscoelastic structure exhibits a non-linear response to the strain rate, in which cyclic stress is out-of-phase with strain, as some of the stored energy is recovered upon removal of the load, while the remaining energy is dissipated as heat. The modulus is represented by a complex quantity: on the one hand the stiffness is defined by elastic behaviour and, on the other hand, the energy dissipative ability of the material is defined by the material’s viscous behaviour. Hence, one could thus hypothesise that the palm stem could be neither an elastic nor a viscous structure, but a combination of both.So now, the aforementioned observations lead us to the following:The herein postulated model envisages the palm stem as a viscoelastic and hollow cylinder prone to Euler and Brazier buckling and ovalization and kinking. This hypothetical model could graphically be imagined as a hollow foam pool noodle with a triple-helicoidal embedded mesh of tough (a high tensile strength) fibre bundles. Both the foam of the pool noodle and the mesh of fibres contribute to the damping while the latter also adds flexural stiffness under bending. The cylinder exhibites a non-linear response to the strain rate, in which cyclic stress is out-of-phase with strain, which makes the whole structure viscoelastic. This envisaging was the main reason why Eq. (10) for Brazier buckling, with a constant t/R for all wind speeds, was experimentally applied to simulate FI of the cocostem of Ref.32.However, it would also be prone to delamination, splitting and shear as the bundles are glued together with “foam” along their screw-like path. The momentum the cylinder should withstand should be a result of dynamic wind loads, mass and inertia that cause a non-linear deformation and pronounced curvature of the cylinder (non-linear due to the varying material properties along the (tapered) stem and structural damping). Strains in the stem would then not be linearly proportional to the load, by which Hooke’s law (ut tensio sic vis) would not be not applicable. And stress would rise non-linearly along the stem radius from the core to the periphery. Progressive fatigue of the wood, or at structural defects (e.g. crack initiations and progressive propagation due to repeated dynamic wind loading), should be taken into account. This model would now possibly align quite well with the scientific findings cited in this paper.Nevertheless, a simple mind experiment can reveal the additional baffling challenges found in real palms: imagine a date palm trunk that has been severely tunneled by Red Palm Weevil and/or pockets of rot: the structure resembles a piece of Gruyère cheese and allows remaining bundles of sound strands to be torn off by hand, as the stiff vascular bundles are just lightly glued together by means of a foamy parenchymatous tissue. The remaining bundles and volumes of sound wood, bordering the void and decayed spaces, could then resemble irregularly shaped columns. Now, imagine the loading of this disk of “Gruyère cheese” due to a bending moment: an infinite variety of kinds of structural failure would take place within the remaining “columns”: buckling, sliding, shear, sideways kinking of the fibres, torsion, crack propagations along a triple-helical path, stress concentrations, et cetera. And as the smart reader will surely agree to, this three-dimensional failure process is totally impossible to depict, or assess, by means of drilling, tomography or simple beam theory. Doubtful readers can have a look at the Figs. 10 and 11 in Ref.17 and imagine that the wood blocks in those figures were the remaining “columns” of our imagined trunk. And, as it can be seen in those figures, the blocks structurally failed due to shear, even under pure axial compression and tension17. And now let us add the following: looping movements of a tall palm in winds has already been recorded18. These looping and circular motions of the stem, inevitably, cause a rotative loading of the cross-section of that same stem. This rotative motion thus causes compression stress (and tension stress on the opposite side) at the periphery and in a circular motion, Real wind loading of palms is thus very dissimilar to the unidirectional loadings (assumed or performed) by e.g.1,3,16,17,18,32,36. And let us add too, that shearing behaviours can be caused due to structural defects (e.g. see Fig. 1), and couple this with the rotative motion and possible progressive fatigue processes in the root system and stem. Now the abovementioned reasonings leave us with a mind-boggling panorama of infinite variables, which seemingly precludes all herein investigated methods from being reliable. However, and from a constructive point of view, these postulated ideas are possibly the best starting point for the development of a future risk assessment method. And the herein offered observations can be used by arboricultural professionals to enhance their tree and palm risk assessment consultancy reports.This is only a partial theory, which need not cancel out others per sé, but may overlap others so as to reach a more acceptable degree of predictive accuracy. This may be a step toward a more complete, fully-unified and more reliable theory that would enable us to make predictions that agree with observations to an acceptable degree of accuracy. Constructing a complete theory from scratch looks excruciatingly difficult now, so perhaps the way forward would be to overlap existing partial theories. Partial theories describe a limited variety of events while leaving others aside. Current partial theories in arboriculture do not seem to be valid on their own19,20,33. Examples are theories that neglect common mechanical behaviours of the wooden body33, simple beam theory and dubious t/R criteria for palm risk assessments. Or predictions of uprooting and breakage that are based on a static wind load analysis, if the latter does not take into account the influence of slenderness, dynamics, mass and inertia in slender and top-loaded (due to e.g. a lion-tailed crown or heavy crop) palms and trees20. A complete theory would thus contain a number of parameters which values, in real-life, cannot be predicted yet and such values may have to be chosen to fit in through experiment. A very appealing goal would be now to overcome this mind-boggling and infinite combination of behaviours and (structural and material) properties, and distill it all into one simple and generic law/model, as was elegantly done for buckling by Ref.24.Researchers have taken sound stems as a starting point (e.g.17,18,32. But, perhaps structurally-damaged trunks should be the place from where to start, as the latter are generally the aim and goal of risk assessments. Future methods could thus perhaps focus on deformations of the stem under circular (wind or artificial) loading, while three-dimensional mechanical behaviours and failures can reasonably be expected within a damaged stem. And also three-dimensional material properties should be taken into account: i.e. MOR and MOE in all anatomical directions. But, as taking those values from published tables would not be feasible (due to the high variability of those properties), different methods from the ones used by e.g.1,3,5,16,17,18,32,36 should perhaps be devised. For instance, a preliminary investigation was carried out on forced vibrations, and resulting resonance frequency values, for a Mexican fan palm, in the light of the identification of trunk decay and its level of severity76. And this could perhaps open up new leads for research. Vibration analysis could monitor repetitive motion signals, to detect abnormal vibration patterns and levels, which could allow the assessment of the overall structural condition of the trunk. But then one would still be left wondering whether that approach would reliably assess e.g. the risk of delamination and crack propagation, or ovalization and kinking.Nevertheless, it is now clear that if we stay within the limits of the theories that are the basis of methods such as e.g. the tree-statics of3, t/R rules used by Ref.1,15 or the ill-fated pulling tests as reported by Ref.77, then our mind will possibly not be able to devise the path of evolution. More

1.Kaneda, T. PRB’s 2016 World Population Data Sheet. http://www.prb.org/Publications/Datasheets/2016/2016-world-population-data-sheet.aspx (2016).2.Bedford, J. et al. A new twenty-first century science for effective epidemic response. Nature 575(7781), 130–136 (2019).CAS 
Article 
ADS 

Google Scholar 
3.Neiderud, C.-J. How urbanization affects the epidemiology of emerging infectious diseases. Infect. Ecol. Epidemiol. 5(1), 27060 (2015).PubMed 

Google Scholar 
4.Dye, C. Health and urban living. Science 319(5864), 766–769 (2008).CAS 
Article 
ADS 

Google Scholar 
5.Fang, C. & Yu, D. Urban agglomeration: An evolving concept of an emerging phenomenon. Landsc. Urban Plan. 162, 126–136 (2017).Article 

Google Scholar 
6.Lee, V. J. et al. Epidemic preparedness in urban settings: New challenges and opportunities. Lancet. Infect. Dis 20(5), 527–529 (2020).CAS 
Article 

Google Scholar 
7.Mollalo, A., Vahedi, B. & Rivera, K. M. GIS-based spatial modeling of COVID-19 incidence rate in the continental United States. Sci. Tot. Environ. 728, 138884 (2020).CAS 
Article 

Google Scholar 
8.Small, C., MacDonald, A. J., & Sousa, D. Spatial network connectivity of population and development in the USA; Implications for disease transmission. Preprint at https://arxiv.org/abs/2004.14237v1 (2020).9.Wang, Z. et al. Quantifying uncertainties in nighttime light retrievals from Suomi-NPP and NOAA-20 VIIRS day/night band data. Remote Sens. Environ. 263, 112557 (2021).Article 
ADS 

Google Scholar 
10.Liu, Q. et al. Spatiotemporal patterns of COVID-19 impact on human activities and environment in Mainland China using nighttime light and air quality data. Remote Sens. 12(10), 1576 (2020).Article 
ADS 

Google Scholar 
11.Elvidge, C., Ghosh, T., Hsu, F.-C., Zhizhin, M. & Bazilian, M. The Dimming of lights in China during the COVID-19 pandemic. Remote Sens. 12(17), 2851 (2020).Article 
ADS 

Google Scholar 
12.Venter, Z.S., Barton, D.N., Gundersen, V., Figari, H., & Nowell, M. Urban nature in a time of crisis: Recreational use of green space increases during the COVID-19 outbreak in Oslo, Norway. Preprint at https://doi.org/10.31235/osf.io/kbdum (2020).13.Chauhan, A. & Singh, R. P. Decline in PM2.5 concentrations over major cities around the world associated with COVID-19. Environ. Res. 187, 109634 (2020).CAS 
Article 

Google Scholar 
14.Kasturi, D. K., Kamarul Zaman, N. A. F., Kaskaoutis, D. G. & Latif, M. T. COVID-19’s impact on the atmospheric environment in the Southeast Asia region. Sci. Tot. Environ. 736, 139658 (2020).Article 

Google Scholar 
15.Forman, R.T.T. Land Mosaics: The Ecology of Landscapes and Regions. (Cambridge University Press, 1995)16.Turner, M. G., Gardner, R. H. & O’Neill, R. V. Landscape Ecology in Theory and Practice (Springer, 2001).
Google Scholar 
17.Wiens, J. A., Schooley, R. L. & Weeks, R. D. Patchy landscapes and animal movements: Do beetles percolate?. Oikos 78(2), 257–264 (1997).Article 

Google Scholar 
18.Dolman, P. M., Hinsley, S. A., Bellamy, P. E. & Watts, K. Woodland birds in patchy landscapes: the evidence base for strategic networks. Ibis 149, 146–160 (2007).Article 

Google Scholar 
19.Laiolo, P. & Tella, J. L. Landscape bioacoustics allow detection of the effects of habitat patchiness on population structure. Ecology 87(5), 1203–1214 (2006).Article 

Google Scholar 
20.Jia, Y., Tang, L., Xu, M. & Yang, X. Landscape pattern indices for evaluating urban spatial morphology—A case study of Chinese cities. Ecol. Ind. 99, 27–37 (2019).Article 

Google Scholar 
21.Bosch, M. & Chenal, J. Spatiotemporal patterns of urbanization in three Swiss urban agglomerations: Insights from landscape metrics, growth modes and fractal analysis. Preprint https://doi.org/10.1101/645549 (2019).Article 

Google Scholar 
22.Wang, H., Huang, Y., Wang, D. & Chen, H. Effects of urban built-up patches on native plants in subtropical landscapes with ecological thresholds—A case study of Chongqing city. Ecol. Indic. 108, 105751 (2020).Article 

Google Scholar 
23.McGarigal, K., & Cushman, S.A. The gradient concept of landscape structure. in Issues and Perspectives in Landscape Ecology (eds. Wiens, J., & Moss, M.) 112–119 (Cambridge University Press, 2005).24.Cushman, S. A., Gutzweiler, K., Evans, J., McGarigal, K. The gradient paradigm: a conceptual and analytical framework for landscape ecology. in Spatial Complexity, Informatics, and Wildlife Conservation (eds. Cushman, S. A. & Huettmann, F.) 83–108. (Springer, 2010).25.McGarigal, K., Tagil, S. & Cushman, S. A. Surface metrics: An alternative to patch metrics for the quantification of landscape structure. Landsc. Ecol. 24(3), 433–450 (2009).Article 

Google Scholar 
26.Fan, C., Myint, S. W., Rey, S. J. & Li, W. Time series evaluation of landscape dynamics using annual Landsat imagery and spatial statistical modeling: Evidence from the Phoenix metropolitan region. Int. J. Appl. Earth Obs. Geoinf. 58, 12–25 (2017).Article 
ADS 

Google Scholar 
27.Kowe, P., Mutanga, O., Odindi, J. & Dube, T. A quantitative framework for analysing long term spatial clustering and vegetation fragmentation in an urban landscape using multi-temporal landsat data. Int. J. Appl. Earth Obs. Geoinf. 88, 102057 (2020).Article 

Google Scholar 
28.Wang, J., Kuffer, M. & Pfeffer, K. The role of spatial heterogeneity in detecting urban slums. Comput. Environ. Urban Syst. 73, 95–107 (2019).Article 

Google Scholar 
29.Wu, D. Q. et al. Multi-scale identification of urban landscape structure based on two-dimensional wavelet analysis: The case of metropolitan Beijing, China. Ecol. Complex. 43, 100832 (2020).Article 

Google Scholar 
30.Rocchini, D. et al. Fourier transforms for detecting multitemporal landscape fragmentation by remote sensing. Int. J. Remote Sens. 34(24), 8907–8916 (2013).Article 

Google Scholar 
31.QGIS Development Team. QGIS Version 3.16.10. Geographic Information System. Open-Source Geospatial Foundation Project. https://www.qgis.org/en/site/ (2019).32.Bennett, M. M. & Smith, L. C. Advances in using multitemporal night-time lights satellite imagery to detect, estimate, and monitor socioeconomic dynamics. Remote Sens. Environ. 192, 176–197 (2017).Article 
ADS 

Google Scholar 
33.Zhuo, L. et al. Modelling the population density of China at the pixel level based on DMSP/OLS non-radiance-calibrated night-time light images. Int. J. Remote Sens. 30(4), 1003–1018 (2009).Article 

Google Scholar 
34.Zhang, Q. & Seto, K. C. Mapping urbanization dynamics at regional and global scales using multi-temporal DMSP/OLS nighttime light data. Remote Sens. Environ. 115(9), 2320–2329 (2011).CAS 
Article 
ADS 

Google Scholar 
35.Yang, B. et al. A spatio-temporal method for crime prediction using historical crime data and transitional zones identified from nightlight imagery. Int. J. Geogr. Inf. Sci. 34(9), 1740–1764 (2020).Article 

Google Scholar 
36.Chen, X. Nighttime lights and population migration: Revisiting classic demographic perspectives with an analysis of recent European data. Remote Sens. 12(1), 169 (2020).Article 

Google Scholar 
37.Li, X., Li, D., Xu, H. & Wu, C. Intercalibration between DMSP/OLS and VIIRS night-time light images to evaluate city light dynamics of Syria’s major human settlement during Syrian Civil War. Int. J. Remote Sens. 38(21), 5934–5951 (2017).Article 

Google Scholar 
38.Duan, X., Hu, Q., Zhao, P., Wang, S. & Ai, M. An approach of identifying and extracting urban commercial areas using the nighttime lights satellite imagery. Remote Sens. 12(6), 1029 (2020).Article 
ADS 

Google Scholar 
39.Krikigianni, E., Tsiakos, C. & Chalkias, C. Estimating the relationship between touristic activities and night light emissions. Eur. J. Remote Sens. 52(sup1), 233–246 (2019).Article 

Google Scholar 
40.Israeli Ministry of Health. Online Geodatabase on COVID-19 Exposures (In Hebrew). https://imoh.maps.arcgis.com/apps/webappviewer/index.html?id=20ded58639ff4d47a2e2e36af464c36e&locale=he&/41.Sharav, N. et al. Gush Dan Metro. The Economic, Social and Urban Impacts of the Gush Dan Metro Network Executive Report. Ministry of Finance, Ministry of Transport and Road Safety, National Economic Council Prime Minister’s Office, NATA, Ayalon Highways, July 2020. (in Hebrew). https://www.gov.il/BlobFolder/reports/metro_goshdan_jul_2020/he/Metro_Executive_Summary_14_july_2020.pdf42.MathWorks 2019. MATLAB ver. 2019b Computer Program. https://www.mathworks.com/. (The MathWorks Inc., 2019). 43.Moellering, H. & Tobler, W. Geographical variances. Geogr. Anal. 4(1), 34–50 (1972).Article 

Google Scholar 
44.Wu, J., Jelinski, D., Luck, M. & Tueller, P. T. Multiscale analysis of landscape heterogeneity: Scale variance and pattern metrics. Ann. GIS 6(1), 6–19 (2000).Article 

Google Scholar 
45.Shen, W., Darrel Jenerette, G., Wu, J. & Gardner, R. H. Evaluating empirical scaling relations of pattern metrics with simulated landscapes. Ecography 27(4), 459–469 (2004).Article 

Google Scholar 
46.Wu, J. Effects of changing scale on landscape pattern analysis: Scaling relations. Landsc. Ecol. 19(2), 125–138 (2004).Article 

Google Scholar 
47.Mahadevan, A. & Campbell, J. W. Biogeochemical patchiness at the sea surface. Geophys. Res. Lett. 29(19), 32-1-32–4 (2002).Article 
ADS 

Google Scholar 
48.Yamamoto, D. Scales of regional income disparities in the USA, 1955 2003. J. Econ. Geogr. 8(1), 79–103 (2007).Article 

Google Scholar 
49.USGS Earth Explorer. https://earthexplorer.usgs.gov.50.Román, M. O. et al. NASA’s black marble nighttime lights product suite. Remote Sens. Environ. 210, 113–143 (2018).Article 
ADS 

Google Scholar 
51.Google Earth Engine Data Catalog, Earth Observation Group, Payne Institute for Public Policy, Colorado School of Mines. https://developers.google.com/earth-engine/datasets/catalog/NOAA_VIIRS_DNB_MONTHLY_V1_VCMCFG. More

Conducting an ecosystem survey during a pandemicCancellation of the survey aboard its primary National Oceanic and Atmospheric Administration (NOAA) survey vessel was overcome through acquisition of a charter for a commercial fishing vessel, following all COVID-19 guidelines (Supplementary Figs. 1 and 2). Initial plans were for 15 days at sea, rather than the 45 typically conducted. This lower effort, along with adverse weather and vessel constraints, resulted in only 25% of the average number of mid-water trawls being collected in the long-term core survey area (Fig. 1 and Supplementary Fig. 1). Despite the data reduction, this effort was one of the only fisheries independent surveys to occur on the US West Coast after the first lockdown in March 2020, furthering the need to evaluate impacts of reduced sampling and provide a robust synthesis of survey results for fishery management. Here we provide updated indices for a selection of ecologically and commercially important species that are critical for assessing ecosystem status.The 2020 sampling was spatially biased towards inshore (shallow) stations (Fig. 1) and thus the previously used method for calculating abundance indices (averaging log-transformed catch-per-unit-effort (CPUE), across all sampled stations) was expected to result in biased indices, in particular for species with strong nearshore (e.g., market squid Dorytheuthis opalescens, anchovy) or offshore (YOY Pacific hake Merluccius productus, myctophids Myctophidae, octopus Octopoda, krill) habitat associations (Supplementary Fig. 3). We confirmed that this bias does indeed occur by recomputing indices for the past 30 years, but using only 1 trawl from each of the 15 stations that were sampled in 2020, and comparing these indices to those using all available trawls (Fig. 2 and Supplementary Fig. 4). In contrast, model-based indices computed from equivalently subsampled past data did not show systematic bias due to the incorporation of spatial covariates (Fig. 2). Thus, although the average log CPUEs were well correlated with model-based indices for well-sampled years (1990–2019), average log CPUEs were determined to be inappropriate for 2020 reporting, and the model-based results were used to develop indices for all taxa for years 1990–2020.Fig. 2: A model for uncertainty and unavoidable effort reduction.a SE of log index vs. number of hauls for a given year from the delta-GLM model. Each point is a year, with 2020 indicated in red. Lines are predicted relationship between SE and sample size for each year, color indicating the mean log index for that year, scaled within taxa. b Relative bias in the index point estimate using 15 hauls from the 2020 stations vs. all hauls from all stations sampled in a given year, computed as (x2020 − xall)/xall. Boxplots show spread of results across all years, 1990–2019 (n = 30 independent years, center: median, box: first and third quartiles, whiskers: smallest and largest values no further than 1.5× IQR from the first and third quartiles; IQR, interquartile range). In the left panel, the index was computed by averaging values of log(CPUE + 1) from all available hauls in a given year. In the right panel, the index was computed from the maximum likelihood estimate (MLE) of a delta-GLM model with spatial covariates, as log(MLE + 1). For the model-based index, the x2020 estimate excludes hauls from the focal year but includes complete data from all other years. CPUE, catch-per-unit-effort; GLM, generalized linear model.Full size imageThe 2020 model-based indices for total rockfish and sanddab (Citharichthys spp.) were the second lowest on record and continued a decline from record high abundance levels observed during the 2014–2016 marine heatwave (Fig. 1)22,23. Pacific hake, myctophids, and octopus were also below average. In contrast, the 2020 index for adult northern anchovy continued a multi-year period of persistently high abundance (Fig. 1). Market squid indices were below average, following a mostly positive trend over the past 7 years. Following the steep decline in 2019, the krill index in 2020 was lower than average (Fig. 1); however, as discussed below, uncertainty may be underestimated for this highly patchy taxonomic group. As a consequence of the low sample sizes, a more rigorous evaluation of the trade-off between sample size (trawls) and uncertainty was conducted, as well as further evaluation of trends through application of existing ecosystem science tools.Quantifying uncertainty by resampling the pastFor most taxa, the uncertainty associated with the 2020 relative abundance estimate was the greatest in the time series, an intuitive result of the sparse sampling for that year (Figs. 1b and 2). The SE was estimated to be over three times the long-term average SE for rockfish and Pacific hake, myctophids, and octopus, and the largest (but less than double the long-term mean) for sanddabs and krill (Fig. 2a). By contrast, the uncertainty associated with the adult anchovy index was lower than the long-term average, due to the great abundance and high frequency of occurrence of anchovy in 2020, compared to years in past decades. This reflects the general trend of uncertainty (on the log scale) being greater for a given taxon when abundance is lower, which generally held for all taxa except krill in our explorations (Fig. 2 and Supplementary Fig. 5). Through time, the relative bias of the subset of stations (2020) vs. the full sample size is also consistently lower for the model-based solution compared to using the average estimate (Fig. 2 and Supplementary Fig. 6). There is also a strong relationship between the number of trawls conducted and the resulting error for each point estimate, with the error essentially doubling when the number of trawls is reduced from the long-term average of 62 to the 15 that were conducted in 2020 (Fig. 2a). By contrast, reducing the total number of trawls from 62 to 40 increases the relative error by just under 25%, while increasing the number of trawls from 62 to 90 only decreases the relative error by 16%. The extent to which the mean relative abundance scales that error up or down, regardless of sample size, is taxon specific. There is an approximate doubling of the error at lowest abundance levels relative to the highest levels for rockfish, sanddabs, hake, and market squid, an increase of more than fourfold over the same range for anchovies and octopus, and relatively modest scaling of the error for myctophids and krill (Fig. 2). This trade-off between survey effort and the error of the ecosystem indices provides critical guidance for future survey planning with respect to the complex trade-off between effort and uncertainty in the face of highly variable interannual catch rates.A seabird’s perspectiveThe Farallon Islands (National Wildlife Refuge) are located in the center of the survey region and host the largest breeding colony of common murre (Uria aalge) in the region (Fig. 1). Interannual variability of Farallon Island seabird population dynamics, reproduction, and foraging ecology are well understood and also track RREAS observations6,17. In particular, patterns such as alternating cycles of forage species occurrence and subsequent reproductive output are known to be linked to regional ocean and climate conditions17,20. Long-term observations of seabird diets in the Farallon Islands were fortunately not impacted by the pandemic. As common murre feed their chicks predominantly either juvenile rockfish or northern anchovy (Supplementary Fig. 7), and common murre prey selection is known to covary with prey abundance in the surrounding ecosystem17,20, these observations provide a critical data stream for evaluating 2020 rockfish and anchovy abundance index estimates from the limited trawl sampling. We updated regression models relating the proportion of rockfish and anchovy in murre diets, respectively, to model-based abundance indices for rockfish and anchovy using past data (Fig. 3). Linear models provided the best fit for YOY rockfish and anchovy, (r2 = 0.70; r2 = 0.58, respectively, both p  More

1.Drucker, P. Indians of the Northwest Coast (McGraw-Hill, 1955).Book 

Google Scholar 
2.Kroeber, A. Culture and natural areas of Native North America (University of California Press, 1939).
Google Scholar 
3.Introduction, S. W. In Handbook of North American Indians volume 7: Northwest Coast (ed. Suttles, W.) 1–15 (Smithsonian Institution, 1990).
Google Scholar 
4.Suttles, W. Coping with abundance: subsistence on the Northwest Coast. In Coast Salish Essays (ed. Suttles, W.) (Talon Books, 1987).
Google Scholar 
5.Barnett, H. G. The Coast Salish of British Columbia (University of Oregon Press, 1955).
Google Scholar 
6.Ames, K. The Northwest Coast: Complex hunter-gatherers, ecology, and social evolution. Annu. Rev. Anthropol. 23, 209–229 (1994).Article 

Google Scholar 
7.Carlson, R. L., Szpak, P. & Richards, M. The Pender Canal site and the beginnings of the Northwest Coast cultural system. Can. J. Archaeol. 41, 1–29 (2017).
Google Scholar 
8.Cannon, A. & Yang, D. Y. Early storage and sedentism on the Pacific Northwest Coast: ancient DNA analysis of salmon remains from Namu, British Columbia. Am. Antiquity 71, 123–140 (2006).Article 

Google Scholar 
9.Matson, R. G. The evolution of Northwest Coast subsistence. In Research in Economic Anthropology Supplement 6: Long-Term Subsistence Change in Prehistoric North America (eds Croes, D. et al.) 366–428 (JAI Press Inc., 1992).
Google Scholar 
10.Caldwell, M. et al. A bird’s eye view of northern Coast Salish intertidal resource management features, southern British Columbia. J. Island Coast. Archaeol. 7, 219–233 (2012).Article 

Google Scholar 
11.Caldwell, M. & Lepofsky, D. Indigenous marine resource management on the Northwest Coast of North America. Ecol. Process. 2(1), 12 (2013).
Google Scholar 
12.Croes, D. R. (ed.). The Qwu?gwes Archaeological Site and Fish Trap (45TN240), and Tested Homestead (45TN396), Eleven-year South Puget Sound Community College Summer Field School Investigations with the Squaxin Island Tribe—Final Report. Report on file, Washington State Department of Archaeology and Historic Preservation, Olympia (2013).13.Lepofsky, D. et al. Shellfish mariculture on the Northwest Coast of North America. Am. Antiq. 80, 236–259 (2015).Article 

Google Scholar 
14.Mathews, D. L. & Turner, N. J. Ocean cultures: northwest coast ecosystems and indigenous management systems. In Conservation for the Anthropocene Ocean: Interdisciplinary Science in Support of Nature and People (eds Levin, P. S. & Poe, M. R.) 169–201 (Academic Press, 2017).Chapter 

Google Scholar 
15.Williams, J. Clam gardens: aboriginal mariculture on Canada’s West Coast (New Star Books, 2006).
Google Scholar 
16.Campbell, S. & Butler, V. Archaeological evidence for resilience of Pacific Northwest salmon populations and the socioecological system over the last~7,500 years. Ecol. Soc. 15(1), 17 (2000).Article 

Google Scholar 
17.Thornton, T., Deur, D. & Kitka, H. Cultivation of salmon and other marine resources on the Northwest Coast of North America. Hum. Ecol. 43, 189–199 (2015).Article 

Google Scholar 
18.Thornton, T. The ideology and practice of Pacific Herring cultivation among the Tlingit and Haida. Hum. Ecol. 43, 213–223 (2015).Article 

Google Scholar 
19.Petersen, J. R. et al. Use of the traditional halibut hook of the Makah Tribe, the čibu⋅d, reduces bycatch in recreational halibut fisheries. PeerJ 8, e9288 (2020).PubMed 
PubMed Central 
Article 

Google Scholar 
20.Ritchie, M. & Angelbeck, B. “Coyote broke the dams”: Power, reciprocity, and conflict in fish weir narratives and implications for traditional and contemporary fisheries. Ethnohistory 67(2), 191–220 (2020).Article 

Google Scholar 
21.Royle, T. C. A. et al. An efficient and reliable DNA-based sex identification method for archaeological Pacific salmonid (Oncorhynchus spp.) remains. PLoS ONE 13(3), e0193212 (2018).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
22.Royle, T. C. A. et al. Investigating the sex-selectivity of a Middle Ontario Iroquoian Atlantic salmon (Salmo salar) and lake trout (Salvelinus namaycush) fishery through ancient DNA analysis. J. Archaeol. Sci. Rep. 31, 102301 (2020).
Google Scholar 
23.George, G. National Energy Board Hearing Order OH-001-2014. Trans Mountain Pipeline ULC. Trans Mountain Expansion Project. Volume 6 (2014).24.Morin, J. Tsleil-Waututh Nation’s History, Culture and Aboriginal Interests in Eastern Burrard Inlet. Report on file, Gowlings, Lafleur, Henderson LLP, Vancouver (2015).25.Suttles, W. Central Coast Salish. In Handbook of North American Indians Volume 7: Northwest Coast (ed. Suttles, W.) 453–475 (Smithsonian Institution, 1990).26.Hancock, M. J. & Marshall, D.E. Catalogue of Salmon Streams and Spawning Escapements of Statistical Area 28 Howe Sound-Burrard Inlet. Canadian Data Report of Fisheries and Aquatic Sciences No. 557 (1986).27.Harris, G. The salmon and trout streams of Vancouver. Waters J. Vanc. Aquar. 3, 4–23 (1978).
Google Scholar 
28.Ricker, W. E. Effects of the Fishery and of Obstacles to Migration on the Abundance of Fraser River Sockeye Salmon (Oncorhynchus nerka). Canadian Technical Report of Fisheries and Aquatic Sciences No. 1522 (1987).29.Charlton, A. S. The Belcarra Park Site. (Department of Archaeology, Simon Fraser University, 1980).30.Lepofsky, D., Trost, D. & Morin, J. Coast Salish interaction: a view from the inlets. Can. J. Archaeol. 31, 190–223 (2007).
Google Scholar 
31.Morin, J., Lepofsky, D., Ritchie, M., Porcic, M. & Edinborough, K. Assessing continuity in the ancestral territory of the Tsleil-Waututh-Coast Salish, southwest British Columbia, Canada. J. Anthropol. Archaeol. 51, 77–87 (2018).Article 

Google Scholar 
32.Morin, J., Muir, B., Ritchie, M. & Sellers, I. Tsleil-Waututh and Simon Fraser University Archaeological Investigations at Port Moody (Reed Point, Shoreline Park, Old Orchard Park, Slaughterhouse Creek, Carraholly Point, and Barnet Beach). Permit 2014–344. Report on file, British Columbia Archaeology Branch, Victoria (2020).33.Harris, C. Voices of disaster: smallpox around the Strait of Georgia in 1782. Ethnohistory 41, 591–626 (1994).Article 

Google Scholar 
34.Chisholm, B. S. Reconstructions of Prehistoric Diet in British Columbia Using Stable-Carbon Isotopic Analysis. PhD dissertation. (Simon Fraser University, 1986).35.Hanson, D. K. Late Prehistoric Subsistence in the Strait of Georgia Region of the Northwest Coast. Master’s thesis. (Simon Fraser University, 1991).36.Trost, T. Forgotten Waters: A Zooarchaeological Analysis of the Cove Cliff Site (DhRr 18), Indian Arm, British Columbia. Master’s thesis. (Simon Fraser University, 2005).37.Pierson, N. Bridging Troubled Waters: Zooarchaeology and Marine Conservation on Burrard Inlet, Southwest British Columbia. Master’s thesis. (Simon Fraser University, 2011).38.Morin, J. et al. DNA-based species identification of ancient salmonid remains provides new insight into pre-contact Coast Salish salmon fisheries in Burrard Inlet, British Columbia, Canada. J. Archaeol. Sci. Rep. 37, 102956 (2021).
Google Scholar 
39.Sproat, G. June 15, 1877. Copy of minute of decision, Joint Indian Reserve Commission. Signed by Dominion Commissioner Alex Anderson, Provincial Commissioner Arch. McKinley and Joint Commissioner G.M. Sproat. Federal set of JIRC’s Minutes and plans, surveyor’s copy. Aboriginal and Northern Affairs Canada, BC Regional Office Specific Claims Branch, Resource Library, Vancouver. AAND Lands and Trusts registration #15215. Also LAC, RG10, Volume 3612, File 3756-23, Reel C10106 (1877).40.Mortimer, H. & George, D. You Call Me Chief: Impressions of the Life of Dan George (Doubleday, 1981).
Google Scholar 
41.Talbot, M. Old Legends and Customs of the British Columbia Coast Indians. s.n., New Westminster (1952).42.Thornton, M. Indian Lives and Legends (Mitchell Press, 1966).
Google Scholar 
43.MacDonald, C., Drake, D., Doerksen, J. & Cotton, M. Between Forest and Sea: Memories of Belcarra (Belcarra Historical Group, 1998).
Google Scholar 
44.Romanoff, S. Fraser Lillooet Fishing. In A Complex Culture of the British Columbia Plateau, Vancouver (ed. Hayden, B.) 222–265 (University of British Columbia Press, 1992).
Google Scholar 
45.Kennedy, D. & Bouchard, R. Sliammon Life, Sliammon Lands (Talonbooks, 1983).
Google Scholar 
46.Mathisen, O. A. The effect of altered sex ratios on the spawning of red salmon. In Studies of Alaska Red Salmon (ed. Koo, T.) 137–246 (University of Washington Press, 1962).
Google Scholar 
47.Reed, W. J. Sex-selective harvesting of Pacific salmon: a theoretically optimal solution. Ecol. Model. 14, 261–271 (1982).Article 

Google Scholar 
48.Salo, E. O. Life history of chum salmon (Oncorhynchus keta). In Pacific Salmon Life Histories (eds Margolis Groot, C. & Margolis, L.) 231–310 (UBC Press, 1991).
Google Scholar 
49.Jenness, D. The Faith of a Coast Salish Indian (British Columbia Provincial Museum, 1955).50.Richling, B. (ed.) The W̲SÁNEĆ and Their Neighbours: Diamond Jenness on the Coast Salish of Vancouver Island, 1935 (Rock’s Mills Press, 2016).51.Dale, C. & Natcher, D. C. What is old is new again: the reintroduction of Indigenous fishing technologies in British Columbia. Local Environ. 20(11), 1309–1321 (2015).Article 

Google Scholar 
52.Ritchie, M. P. & Springer, C. Harrison River Chum Fishery: The Ethnographic and Archaeological Perspective. Report on file, Sts′ailes, Agassiz (2010).53.Simeone, W. E. & Valentine, E. M. Ahtna Knowledge of Long-Term Changes in Salmon runs in the Upper Copper River Drainage, Alaska (Alaska Department of Fish and Game, Division of Subsistence, 2007).54.Langdon, S. J. Traditional Knowledge and Harvesting of Salmon by Huna and Hinyaa Tlingit. (U.S. Fish and Wildlife Service, Office of Subsistence Management, Fisheries Resource Monitoring Program, 2006).55.Ratner, N. C. et al. Local Knowledge, Customary Practices, and Harvest of Sockeye salmon from the Klawock and Sarkar Rivers, Prince of Wales Island, Alaska (Alaska Department of Fish and Game, Division of Subsistence, 2006)56.Curtis, E. The North American Indian: being a series of volumes picturing and describing the Indians of the United States, the Dominion of Canada, and Alaska, Vol. 13 (Plimpton Press, 1924).
Google Scholar 
57.Kennedy, D. & Bouchard, R. Stl’atl’imx fishing. In A Complex Culture of the British Columbia Plateau (ed. Hayden, B.) 266–354 (UBC Press, 1992).
Google Scholar 
58.Yang, D. Y. & Watt, K. Contamination controls when preparing archaeological remains for ancient DNA analysis. J. Archaeol. Sci. 32(3), 331–336 (2005).Article 

Google Scholar 
59.Speller, C. F. et al. High potential for using DNA from ancient herring bones to inform modern fisheries management and conservation. PLoS ONE 7, e51122 (2012).CAS 
PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
60.Yang, D. Y., Eng, B., Waye, J. S., Dudar, J. C. & Saunders, S. R. Technical note: improved DNA extraction from ancient bones using silica-based spin columns. Am. J. Phys. Anthropol. 105(4), 539–543 (1998).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
61.Yang, D. Y., Liu, L., Chen, X. & Speller, C. F. Wild or domesticated: DNA analysis of ancient water buffalo remains from North China. J. Archaeol. Sci. 35(10), 2778–2785 (2008).Article 

Google Scholar 
62.Bertho, S. et al. The unusual rainbow trout sex determination gene hijacked the canonical vertebrate gonadal differentiation pathway. Proc. Natl. Acad. Sci. U.S.A. 115(50), 12781–12786 (2008).Article 
CAS 

Google Scholar 
63.Yano, A. et al. An immune-related gene evolved into the master sex-determining gene in rainbow trout, Oncorhynchus mykiss. Curr. Biol. 22(15), 1423–1428 (2012).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
64.Yano, A. et al. The sexually dimorphic on the Y-chromosome gene (sdY) is a conserved male-specific Y-chromosome sequence in many salmonids. Evol. Appl. 6(3), 486–496 (2013).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
65.Yang, D., Cannon, A. & Sanders, S. R. DNA species identification of archaeological salmon bone from the Pacific Northwest Coast of North America. J. Archaeol. Sci. 31, 619–631 (2004).Article 

Google Scholar 
66.R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2019).67.Kim, K. et al. A real-time PCR-based amelogenin Y allele dropout assessment model in gender typing of degraded DNA samples. Int. J. Legal Med. 127, 55–61 (2013).PubMed 
Article 
PubMed Central 

Google Scholar 
68.Sinding, M. et al. Sex determination of baleen whale artefacts: Implications for ancient DNA use in zooarchaeology. J. Archaeol. Sci. Rep. 10, 345–349 (2016).
Google Scholar 
69.Cooper, A. & Poinar, H. Ancient DNA: do it right or not at all. Science 289(5482), 1139 (2000).CAS 
PubMed 
Article 
PubMed Central 

Google Scholar 
70.McKechnie, I. Investigating the complexities of sustainable fishing at a prehistoric village on western Vancouver Island, British Columbia, Canada. J. Nat. Conserv. 15(3), 208–222 (2007).Article 

Google Scholar 
71.Cannon, A., Yang, D. Y. & Speller, C. Site-specific salmon fisheries on the central coast of British Columbia. In The Archaeology of North Pacific Fisheries (eds Moss, M. & Cannon, A.) 57–74 (University of Alaska Press, 2011).
Google Scholar 
72.McKechnie, I. & Moss, M. Meta-analysis in zooarchaeology expands perspectives on Indigenous fisheries of the Northwest Coast of North America. J. Archaeol. Sci. Rep. 8, 470–485 (2016).
Google Scholar 
73.Orchard, T. J. & Szpak, P. Zooarchaeological and isotopic insights into locally variable economic patterns: a case study from late Holocene southern Haida Gwaii, British Columbia. BC Stud. 187, 107–147 (2015).
Google Scholar 
74.Rodrigues, A. T., McKechnie, I. & Yang, D. Y. Ancient DNA analysis of indigenous rockfish use on the Pacific Coast: implications for marine conservation areas and fisheries management. PLoS ONE 13(2), e0192716 (2018).PubMed 
PubMed Central 
Article 
CAS 

Google Scholar 
75.McGregor, D. Coming full circle: Indigenous knowledge, environment, and our future. Am. Indian Q. 28(3), 385–410 (2004).Article 

Google Scholar 
76.Suttles, W. Economic Life of the Coast Salish of Hario and Rosario Straits. PhD Dissertation. (University of Washington, 1951).77.Caldwell, M. E. Northern Coast Salish Marine Resource Management. PhD dissertation. (University of Alberta, 2015).78.Deur, D. Tending the garden, making the soil: Northwest Coast estuarine gardens as engineered environments. In Keeping It Living: Traditions of Plant Use and Cultivation on the Northwest Coast of North America (eds Deur, D. & Turner, N.) (UBC Press, 2005).
Google Scholar 
79.Hoffmann, T. et al. Engineered feature used to enhance gardening at a 3800-year-old site on the Pacific Northwest coast. Sci. Adv. 2(12), e1601282 (2016).PubMed 
PubMed Central 
Article 
ADS 

Google Scholar 
80.Lepofsky, D. et al. Documenting pre-contact plant management on the Northwest Coast: an example of prescribed burning in the Central and Upper Fraser Valley, British Columbia. In Keeping It Living: Traditions of Plant Use and Cultivation on the Northwest Coast of North America (eds Deur, D. & Turner, N.) 218–239 (UBC Press, 2005).
Google Scholar 
81.Turner, N. J., Deur, D. & Lepofsky, D. Plant management systems of British Columbia’s First Peoples. BC Stud. 179, 107–133 (2013).
Google Scholar 
82.Turner, N. J., Smith, R. & Jones, J. A fine line between two nations: ownership patterns for plant resources among Northwest Coast indigenous peoples. In Keeping It Living: Traditions of Plant Use and Cultivation on the Northwest Coast of North America (eds Deur, D. & Turner, N. J.) 151–180 (UBC Press, 2005).
Google Scholar 
83.Turner, N. J. & Peacock, S. Solving the perennial paradox: ethnobotanical evidence for plant resource management on the Northwest Coast. In Keeping It Living: Traditions of Plant Use and Cultivation on the Northwest Coast of North America (eds Deur, D. & Turner, N. J.) 101–151 (UBC Press, 2005).
Google Scholar 
84.Limburg, K. E., Walther, Y., Hong, B., Olson, C. & Stora, J. Prehistoric versus modern Baltic Sea cod fisheries: selectivity across the millennia. Proc. R. Soc. B 275, 2659–2665 (2008).PubMed 
PubMed Central 
Article 

Google Scholar 
85.Sanchez, G. Indigenous stewardship of marine and estuarine fisheries? Reconstructing the ancient size of Pacific herring through linear regression models. J. Archaeol. Sci. Rep. 29, 102061 (2020).
Google Scholar 
86.Slaney, T. L., Hyatt, K. D., Northcote, T. G. & Fielden, R. J. Status of anadromous salmon and trout in British Columbia and Yukon. Fisheries 21(10), 20–35 (1996).Article 

Google Scholar 
87.Kope, R. & Wainwright, T. Trends in the status of Pacific salmon populations in Washington, Oregon, California, and Idaho. N. Pac. Anadr. Fish Comm. Bull. 1, 1–12 (1998).
Google Scholar 
88.Gustafson, R. G. et al. Pacific salmon extinctions: Quantifying lost and remaining diversity. Conserv. Biol. 21(4), 1009–1020 (2007).PubMed 
Article 
PubMed Central 

Google Scholar 
89.Price, M. H., English, K. K., Rosenberger, A. G., MacDuffee, M. & Reynolds, J. D. Canada’s wild salmon policy: an assessment of conservation progress in British Columbia. Can. J. Fish. Aquat. Sci. 74(10), 1507–1518 (2017).Article 

Google Scholar 
90.Gayeski, N. J. et al. The failure of wild salmon management: need for a place-based conceptual foundation. Fisheries 43(7), 303–309 (2018).Article 

Google Scholar 
91.Morales, Q. E., Lepofsky, D. & Berkes, F. Ethnobiology and fisheries: Learning from the past for the present. J. Ethnobiol. 37(3), 369–379 (2017).Article 

Google Scholar 
92.Reid, A. J. et al. Two-eyed seeing: an Indigenous framework to transform fisheries research and management. Fish Fish. 00, 1–19 (2020).
Google Scholar 
93.Atlas, W. I. et al. Indigenous systems of management for culturally and ecologically resilient Pacific salmon (Oncorhynchus spp.) fisheries. Bioscience 71(2), 1–19 (2021).Article 

Google Scholar  More