Philippa Kaur

1.
Mitchard, E. T. A. The tropical forest carbon cycle and climate change. Nature 559, 527–534 (2018).
ADS  CAS  PubMed  Google Scholar 
2.
Saatchi, S. S. et al. Benchmark map of forest carbon stocks in tropical regions across three continents. Proc. Natl Acad. Sci. USA 108, 9899–9904 (2011).
ADS  CAS  PubMed  Google Scholar 

3.
Baccini, A. et al. Estimated carbon dioxide emissions from tropical deforestation improved by carbon-density maps. Nat. Clim. Change 2, 182–185 (2012).
ADS  CAS  Google Scholar 

4.
Liu, Y. Y. et al. Recent reversal in loss of global terrestrial biomass. Nat. Clim. Change 5, 470–474 (2015).
ADS  Google Scholar 

5.
Harris, N. L. et al. Baseline map of carbon emissions from deforestation in tropical regions. Science 336, 1573–1576 (2012).
ADS  CAS  PubMed  Google Scholar 

6.
Marco, M. D., Watson, J. E. M., Currie, D. J., Possingham, H. P. & Venter, O. The extent and predictability of the biodiversity–carbon correlation. Ecol. Lett. 21, 365–375 (2018).
PubMed  Google Scholar 

7.
Giardina, F. et al. Tall Amazonian forests are less sensitive to precipitation variability. Nat. Geosci. 11, 405–409 (2018).
ADS  CAS  Google Scholar 

8.
Erb, K.-H. et al. Unexpectedly large impact of forest management and grazing on global vegetation biomass. Nature 553, 73–76 (2018).
ADS  CAS  PubMed  Google Scholar 

9.
Zarin, D. J. et al. Can carbon emissions from tropical deforestation drop by 50% in 5 years? Glob. Change Biol. 22, 1336–1347 (2016).
ADS  Google Scholar 

10.
Chaplin-Kramer, R. et al. Degradation in carbon stocks near tropical forest edges. Nat. Commun. 6, 10158 (2015).
ADS  CAS  PubMed  PubMed Central  Google Scholar 

11.
Brandt, M. et al. Satellite passive microwaves reveal recent climate-induced carbon losses in African drylands. Nat. Ecol. Evol. 2, 827 (2018).
PubMed  Google Scholar 

12.
Fan, L. et al. Satellite-observed pantropical carbon dynamics. Nat. Plants 5, 944–951 (2019).
CAS  PubMed  Google Scholar 

13.
Mitchard, E. T. et al. Markedly divergent estimates of Amazon forest carbon density from ground plots and satellites. Glob. Ecol. Biogeogr. 23, 935–946 (2014).
PubMed  PubMed Central  Google Scholar 

14.
Mitchard, E. T. et al. Uncertainty in the spatial distribution of tropical forest biomass: a comparison of pan-tropical maps. Carbon Balance Manag. 8, 10 (2013).
PubMed  PubMed Central  Google Scholar 

15.
Avitabile, V. et al. An integrated pan-tropical biomass map using multiple reference datasets. Glob. Change Biol. 22, 1406–1420 (2016).
ADS  Google Scholar 

16.
Réjou-Méchain, M. et al. Upscaling forest biomass from field to satellite measurements: Sources of errors and ways to reduce them. Surv. Geophys. 40, 881–911 (2019).
ADS  Google Scholar 

17.
Saatchi, S. Mapping tropical forest biomass: synthesis of ground and remote sensing inventory. Consult. Rep. 2 High Carbon Stock Sci. Study (2015).

18.
Ploton, P. et al. A map of African humid tropical forest aboveground biomass derived from management inventories. Sci. Data 7, 221 (2020).
PubMed  PubMed Central  Google Scholar 

19.
Philippon, N. et al. The light-deficient climates of Western Central African evergreen forests. Environ. Res. Lett. 14, 034007 (2018).

20.
Saatchi, S. et al. Seeing the forest beyond the trees. Glob. Ecol. Biogeogr. 24, 606–610 (2015).
Google Scholar 

21.
Mermoz, S., Le Toan, T., Villard, L., Réjou-Méchain, M. & Seifert-Granzin, J. Biomass assessment in the Cameroon savanna using ALOS PALSAR data. Remote Sens. Environ. 155, 109–119 (2014).
ADS  Google Scholar 

22.
Lewis, S. L. et al. Above-ground biomass and structure of 260 African tropical forests. Philos. Trans. R. Soc. B Biol. Sci. 368, 20120295 (2013).
Google Scholar 

23.
Hansen, M. C., Potapov, P. & Tyukavina, A. Comment on “Tropical forests are a net carbon source based on aboveground measurements of gain and loss”. Science 363, eaar3629 (2019).
CAS  PubMed  Google Scholar 

24.
Baccini, A. et al. Tropical forests are a net carbon source based on aboveground measurements of gain and loss. Science 358, 230–234 (2017).
ADS  MathSciNet  CAS  PubMed  PubMed Central  MATH  Google Scholar 

25.
Breiman, L. Random forests. Mach. Learn 45, 5–32 (2001).
MATH  Google Scholar 

26.
Lyapustin, A., Wang, Y., Korkin, S. & Huang, D. MODIS Collection 6 MAIAC algorithm. Atmos. Meas. Tech. 11, 5741–5765 (2018).

27.
Fick, S. E. & Hijmans, R. J. WorldClim 2: new 1‐km spatial resolution climate surfaces for global land areas. Int. J. Climatol. 37, 4302–4315 (2017).
Google Scholar 

28.
Kühn, I. Incorporating spatial autocorrelation may invert observed patterns. Divers. Distrib. 13, 66–69 (2007).
Google Scholar 

29.
Dormann, C. F. Effects of incorporating spatial autocorrelation into the analysis of species distribution data. Glob. Ecol. Biogeogr. 16, 129–138 (2007).
Google Scholar 

30.
Roberts, D. R. et al. Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. Ecography 40, 913–929 (2017).
Google Scholar 

31.
Valavi, R., Elith, J., Lahoz‐Monfort, J. J. & Guillera‐Arroita, G. blockCV: an r package for generating spatially or environmentally separated folds for k-fold cross-validation of species distribution models. Methods Ecol. Evol. 10, 225–232 (2019).
Google Scholar 

32.
Parmentier, I. et al. Predicting alpha diversity of African rain forests: models based on climate and satellite-derived data do not perform better than a purely spatial model. J. Biogeogr. 38, 1164–1176 (2011).
Google Scholar 

33.
Baccini, A., Walker, W., Carvalho, L., Farina, M. & Houghton, R. A. Response to Comment on “Tropical forests are a net carbon source based on aboveground measurements of gain and loss”. Science 363, eaat1205 (2019).
CAS  PubMed  MATH  Google Scholar 

34.
Xu, L., Saatchi, S. S., Yang, Y., Yu, Y. & White, L. Performance of non-parametric algorithms for spatial mapping of tropical forest structure. Carbon Balance Manag. 11, 18 (2016).
PubMed  PubMed Central  Google Scholar 

35.
Dormann, C. F. et al. Methods to account for spatial autocorrelation in the analysis of species distributional data: a review. Ecography 30, 609–628 (2007).
Google Scholar 

36.
Irwin, A. The ecologist who wants to map everything. Nature 573, 478–481 (2019).
ADS  PubMed  Google Scholar 

37.
Legendre, P. Spatial autocorrelation: trouble or new paradigm? Ecology 74, 1659–1673 (1993).
Google Scholar 

38.
Meyer, H., Reudenbach, C., Wöllauer, S. & Nauss, T. Importance of spatial predictor variable selection in machine learning applications–Moving from data reproduction to spatial prediction. Ecol. Model. 411, 108815 (2019).
Google Scholar 

39.
Strobl, C., Boulesteix, A.-L., Kneib, T., Augustin, T. & Zeileis, A. Conditional variable importance for random forests. BMC Bioinformatics 9, 307 (2008).
PubMed  PubMed Central  Google Scholar 

40.
Lefsky, M. A. et al. Estimates of forest canopy height and aboveground biomass using ICESat. Geophys. Res. Lett. 32, L22S02 (2005).

41.
Réjou-Méchain, M. et al. Upscaling Forest biomass from field to satellite measurements: sources of errors and ways to reduce them. Surv. Geophys. 40, 881–911 (2019).

42.
Mitchard, E. T. A. et al. Comment on ‘A first map of tropical Africa’s above-ground biomass derived from satellite imagery’. Environ. Res. Lett. 6, 049001 (2011).
ADS  Google Scholar 

43.
Asner, G. P. et al. High-resolution carbon mapping on the million-hectare Island of Hawaii. Front. Ecol. Environ. 9, 434–439 (2011).
Google Scholar 

44.
Asner, G. P. et al. Human and environmental controls over aboveground carbon storage in Madagascar. Carbon Balance Manag. 7, 2 (2012).
PubMed  PubMed Central  Google Scholar 

45.
Asner, G. P. et al. High-resolution mapping of forest carbon stocks in the Colombian Amazon. Biogeosciences 9, 2683 (2012).
ADS  CAS  Google Scholar 

46.
Asner, G. P. et al. Mapped aboveground carbon stocks to advance forest conservation and recovery in Malaysian Borneo. Biol. Conserv. 217, 289–310 (2018).
Google Scholar 

47.
Xu, L. et al. Spatial distribution of carbon stored in forests of the Democratic Republic of Congo. Sci. Rep. 7, 15030 (2017).
ADS  PubMed  PubMed Central  Google Scholar 

48.
Schepaschenko, D. et al. The Forest Observation System, building a global reference dataset for remote sensing of forest biomass. Sci. Data 6, 1–11 (2019).
Google Scholar 

49.
Chave, J. et al. Ground data are essential for biomass remote sensing missions. Surv. Geophys. 40, 863–880 (2019).
ADS  Google Scholar 

50.
van den Hoogen, J. et al. Soil nematode abundance and functional group composition at a global scale. Nature 572, 194–198 (2019).
ADS  CAS  PubMed  Google Scholar 

51.
Steidinger, B. S. et al. Climatic controls of decomposition drive the global biogeography of forest-tree symbioses. Nature 569, 404–408 (2019).
ADS  CAS  PubMed  Google Scholar 

52.
Bastin, J.-F. et al. The global tree restoration potential. Science 365, 76–79 (2019).
ADS  CAS  PubMed  Google Scholar 

53.
Trabucco, A. & Zomer, R. J. Global aridity index (global-aridity) and global potential evapo-transpiration (global-PET) geospatial database. CGIAR Consort Spat Information (2009).

54.
Wilson, A. M. & Jetz, W. Remotely sensed high-resolution global cloud dynamics for predicting ecosystem and biodiversity distributions. PLoS Biol. 14, e1002415 (2016).
PubMed  PubMed Central  Google Scholar 

55.
Bowman, D. M., Williamson, G. J., Keenan, R. J. & Prior, L. D. A warmer world will reduce tree growth in evergreen broadleaf forests: evidence from A ustralian temperate and subtropical eucalypt forests. Glob. Ecol. Biogeogr. 23, 925–934 (2014).
Google Scholar 

56.
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).
ADS  CAS  PubMed  Google Scholar 

57.
Rennó, C. D. et al. HAND, a new terrain descriptor using SRTM-DEM: Mapping terra-firme rainforest environments in Amazonia. Remote Sens. Environ. 112, 3469–3481 (2008).
ADS  Google Scholar 

58.
Nachtergaele, F., Velthuizen, H. V., Verelst, L. & Wiberg, D. Harmonized World Soil Database (HWSD) (Food and Agriculture Organization, U. N. Rome, 2009).

59.
Defourny, P. et al. Algorithm Theoretical Basis Document for Land Cover Climate Change Initiative. Technical report (European Space Agency, 2014).

60.
Segal, M. & Xiao, Y. Multivariate random forests. WIREs Data Min. Knowl. Discov. 1, 80–87 (2011).
Google Scholar 

61.
CCI, ESA. New Release of 300 m Global Land Cover and 150 m Water Products (v.1.6.1) and new version of the User Tool (3.10) for Download (ESA CCI Land cover website, 2016). More

Global P losses from soils and soil P balances
All continents result in negative P balances (e.g., net P losses from agricultural systems, Table 1; Fig. 2) except Asia, Oceania and Australia, with Asia having a slightly positive but near zero P balance. This is in spite of high to very high chemical fertilizer inputs (with a range of 1.7 to 13 kg ha−1 yr−1 between the different continents, with national values reaching up to 14 and 19 kg ha−1 yr−1 for the European Union 15 old Member States (EU15, member states joining before 2004 mainly in western and northern Europe) and China, respectively). Most negative P balances are indicated for Africa due to very low chemical fertilizer input of 1.7 kg ha−1 yr−1 paired with high losses due to soil erosion of 9.6 kg ha−1 yr−1. South America as well as Central and Eastern Europe (NEU11; which are the new member states joining the EU after 2004 with the exception of Cyprus and Malta) also exhibit high P losses but for different reasons. South America has a very high chemical fertilizer input but also high losses due to soil erosion paired with high P exports due to organic P management (calculated as the sum of manure and residue input minus plant uptake). In contrast, the eastern European Union New Member States (NEU11) have rather low erosional losses but also very low chemical fertilizer input. With the hypothetical assumption of no replenishment due to chemical fertilizer (e.g., due to economic or technical constraints), calculation of soil P balances results in negative balances globally, as well as for all the continents and regions considered (depletion between 4 and 20 kg P ha−1 yr−1; Figs. 3 and 4; Table 1). The latter demonstrates the vulnerability of today’s global land management system and its strong dependency on chemical P fertilizers from non-renewable mineable P deposits.
Table 1 Main phosphorus (P) statistics in kg P ha−1yr−1 for all continents and selected countries.
Full size table

Fig. 2: Global average phosphorus (P) losses due to soil erosion in kg ha−1 yr−1.

The chromatic scale represents the P losses estimates, while the gray color indicates the cropland areas that were excluded from the modeling due to data unavailability. Note that classes are not regularly scale ranked but are divided into six classes using the quantile classification method. Only plant available fractions were considered. For the more residual P fractions please refer to Table 1 or Figs. 3 and 4).

Full size image

Fig. 3: Global P soil pools and depletion due to erosion.

Arrows indicate fluxes (positive: net input to soils, negative: depletion of soils). *Organic P management = sum of manure and residue input minus plant uptake. Non-plant P = non-plant available P. Inorganic and organic P give plant available fractions. Total soil P: sum of P fractions lost from soil via erosion with relative errors. No/with chemical = P balance with and without chemical fertilizer.

Full size image

Fig. 4: Soil P pools and depletion due to erosion in Africa, Europe and North America.

AD = Atmospheric Deposition. CF = Chemical Fertilizer. OM = Organic P management = sum of manure and residue input minus plant uptake. Arrows indicate fluxes (positive: net input to soils, negative: depletion of soils). Non-plant P = non-plant available P. Inorganic and organic P give plant available fractions. Soil Plost: sum of P fractions lost from soil via erosion with relative errors. No/with chemical = P balance with and without chemical fertilizer.

Full size image

Our area related calculations result in an average P loss for arable soils, due to erosion by water, of approximately (5.9_{ – 0.79% }^{ + 1.17% }) kg ha−1 yr−1 globally (Fig. 3, Table 1). This is around 60% of the rates given by Smil16 who estimated 10 kg P ha−1 yr−1 from arable fields due to soil erosion by water. Our total P losses due to soil erosion by water from arable soils globally result in 6.3 Tg yr−1 with 1.5 Tg yr−1 for organic and 4.8 Tg yr−1 for inorganic P. With these values the results are at the lower end of the range discussed in the literature (between 1–19 Tg yr−1, Table 2).
Table 2 Global fluxes from soils/arable systems to waters as discussed in recent literature. For comparability, all values were normalized to 1 billion ha of arable land (with original values given in brackets).
Full size table

Liu34 in estimating net input to, and output from, cropland systems indicated a net loss of P from the world’s croplands of about 12.8 Tg  yr−1 (calculating input from atmosphere, weathering and chemical fertilizer versus output from organic P management, soil erosion and runoff), which would be, according to their calculations, the same order of magnitude as synthetic fertilizer input (13.8 Tg yr−1 for statistical year 2003/2004). Our calculations result in an approximate net soil loss due to erosion of 6.3 Tg yr−1 with a global average chemical fertilizer input of 9.2 Tg yr−1. The fluxes clearly show the critical dependence on chemical fertilizer globally, with a hypothetical net-average-area related depletion of 10.7 kg ha−1 yr−1 globally without the compensation due to chemical fertilizer, from which a loss of 5.2 kg ha−1 yr−1 stems from organic P management (sum of manure and residue input minus plant uptake) and 5.9 kg ha−1 yr−1 from soil erosion (Table 1). Continental and national erosional P losses are between 40 and 85% of total P losses from agricultural systems with the exception of Europe and Australia (16 and 19%, respectively). Globally, as well as for Africa, South America and Asia, P soil losses due to erosion are higher than losses due to organic P management.
A recent quantification of atmospheric P dust input, based on dust measurements in the Sierra Nevada, concluded that measured dust fluxes are greater than, or equal to, modern erosional outputs and a large fractional contribution relative to bedrock35. However, even though their measured and modeled maximum atmospheric P fluxes were in the same order of magnitude as the atmospheric flux data of Wang et al.36 used in this study (0.1 kg ha−1 yr−1 for North America), erosion rates from sediment trap measurements in their investigated forest ecosystems were considerably lower (maximum of 0.06 kg ha−1 yr−135) than might be expected in arable lands worldwide. As such, we can clearly not agree with the above conclusion.
Mitigation of the soils P status in the long term by decreasing the deficits of the current organic P management seems difficult and rather unlikely in many regions of the world (see discussion of continental balances below). As such, and considering the expected shortage of P supply from industrial fertilizers in the future, the evaluation of P fluxes clearly shows that soil erosion has to be limited to the feasible absolute minimum in the future.
Only a small fraction of the total soil P is plant available, because a large fraction is either bound in, adsorbed to, or made unavailable by occlusion in minerals (apatite and occluded P)37. Our modeling approach follows the approach of Yang et al.37 that builds on existing knowledge of soil P processes and data bases to provide spatially explicit estimates of different forms of naturally occurring soil P on the global scale. Yang et al.37 uses data acquired with the Hedley fractionation method38 which splits soil P into different fractions that are extracted sequentially with successively stronger reagents and which are merged into various inorganic and organic pools. There is great uncertainty to associate these fractions with functional plant uptake39,40 but some recent work aims at quantifying residence times of the different fractions17,41. We present the total P loss as well as the plant-available pool (Figs. 3 and 4, Table 1) as the most labile P which can participate to plant nutrition in a time scale up to months17 (i.e., the so-called labile inorganic P, inorganic P bound to secondary minerals, labile and stable organic P33,37). We contrast these more labile, short lived fractions with fractions considered very stable which would not be plant available at short time scales (i.e., inorganic P associated with minerals such as apatite or occluded P17, corresponding to 52% of total soil P in our estimates at the global scale).
Verification of soil P loss with a comparison to riverine P exports
Verification of our proposed P soil losses might be done indirectly via a comparison to riverine P loads. However, to verify our on-site P soil loss approach with off-site P river export data requires two prerequisites: (i) an estimation of the agricultural erosion and runoff contribution to total P loads in the rivers (as the latter will be the total sum of agricultural, urban and industrial runoff) and (ii) an assumption on sediment delivery rates (e.g., the percent of sediments reaching the rivers from the total eroded sediments) to estimate P loads to rivers from on-site P soil loss. Regarding the first prerequisite, we searched for published river P export data with a separation of the agricultural erosion and runoff contribution from total P loads in the rivers. Regarding the second prerequisite, there is a lack of models describing integrated sediment-delivery to rivers on a continental or global scale, but it has been discussed that sediment delivery ratios generally decrease as drainage area increases, ranging from roughly 30–100% in small catchments (≤0.1 km2) to 2–20% at large spatial scales (e.g., ≥1000 km2)42. For our comparison, we used a range of average sediment delivery rates between 11–30% as used in recent large scale studies from continental to global scale43,44. In doing so, we would like to point out, that even if 70–89% of the P lost from agricultural soils might be re-deposited within catchments, potential threats of P loss from soils are not reduced. Erosion (and thus P loss) occurs predominantly on agricultural soils while re-deposition will mostly occur in depositional hollows, wetlands, riparian zones or buffer strips. Thus, P is lost as a nutrient on food and feed production sites but re-deposited as a potential ecological threat to biodiversity and ecosystem health due to its eutrophication effect in less intensively or unmanaged ecosystems. Last but not least we would like to point out that RUSLE only considers soil displacement due to rill and inter-rill erosion neither considering tillage and gully erosion nor land sliding.
A comparison of calculated potential P export to rivers from our on-site soil P losses is well within the range of published riverine P exports (Table 3). Beusen et al.45 used total suspended sediment measurements from the GEMS-GLORI database to extrapolate spatially distributed sediment rates for the world’s largest rivers with land use, topography, lithology and precipitation as factors in a multiple linear regression approach accounting for soil erosion as well as sediment trapping. The associated nutrient exports for all continents, as well as global assessment were made by calibrating nutrient export to sediment rates with the model Global News45 using established correlations between sediment and nutrient concentrations. In comparing their P export with suspended sediments in rivers45 to our assessments, we underestimated P export globally as well as for all continents, with the exception of Africa (Table 3), which is, however, (i) strongly related to assumed sediment delivery rates and (ii) our RUSLE application only considering rill and inter-rill erosional processes with an unknown contribution of gullies, landslides and tillage erosion.
Table 3 On-site P loss from gross soil erosion (this study), calculated potential riverine loads with sediment delivery ratios between 11–30%43,44 and comparison to global and regional riverine P export studies. All values in kg P ha−1yr−1.
Full size table

An analysis of 17 large scale European catchments (250–11,000 km2) quantifying the loss of P to surface waters from an off-site perspective (P flux measurement in waters) resulted in 0.05–1.5 kg ha−1 yr−1 of P loss due to soil erosion and runoff from agricultural lands into streams and rivers (excluding a Greek catchment with a very high P export of 6 kg ha−1 yr−1)46. Recalculating our on-site soil P losses with sediment delivery ratios between 11–30% results in rates for geographic Europe between 0.1–0.4 kg ha−1 yr−1 which are at the lower end of the range of Kronvang, et al.46 (Table 3). Potential P losses as riverine export due to agricultural runoff and erosion of 143 watersheds across the U.S.47 are in the same range as the calculated P loss assessment of our study, while P loads to Lake Erie assessed in a regional study48 seem slightly higher (Table 3). However, no partitioning between agricultural, urban or industrial fluxes was possible from the latter study. An assessment of the Yangtze River Basin (with 1.8 Million km2 near 20% of the whole Chinese territory) gives modeled soil P losses (on-site perspective) between 0–196 kg ha−1 yr−1 demonstrating the huge spatial heterogeneity of on-site soil erosion rates49. The model output was calibrated with total measured nutrient loads in rivers, while the partitioning differed between dissolved point and non-point as well as adsorbed non-point pollution which can be mostly attributed to soil erosion of agricultural fields. The average P loads due to soil erosion from agricultural fields in the Yangtze River Basin compares well with the range assessed in our study for China (2.7 versus a range of 1.4 to 3.7 kg ha−1 yr−1, respectively). The same holds true for assessment comparisons of Africa50 and South America51 (Table 3), even though it should be considered that especially for these latter studies scale differs considerably from our approach.
Regional P losses and balances
Parallel to the distribution pattern and dynamics of global soil erosion by water32, P losses from soils due to water erosion are most dramatic in countries and regions with intensive agriculture and/or extreme climates (e.g., droughts followed by significant rain events or high frequencies of heavy rain storms) due to high erosivity effects52. As such, our calculations result in extremely high P losses due to erosion ( >20 kg ha−1 yr−1) in regions such as eastern China, many regions in Indonesia, parts of east and south-eastern Africa (Ethiopia, Eritrea, Mozambique), Central America and parts of South America (South-Eastern Brazil; Southern Chile, Peru (Fig. 2)). A very high P loss (10 to 20 kg ha−1 yr−1) is estimated for parts of Southern Africa (South Africa, Madagascar, Tanzania) and South America (Bolivia) and a high loss (5–10 kg ha−1 yr−1) for most of India, as well as regions in Southern Africa (Angola, Zambia) and South America (Uruguay) (Fig. 2). Even though the underlying erosion model algorithm does not calculate the net catchment output but rather the on-site displacement of soil sediments which might then be re-located to other parts of the fields or even buried at depositional places, the considered on-site field management will clearly be confronted with substantial P losses due to soil erosion by water. Only considering agronomic P inputs and outputs without including P losses due to erosion by MacDonald et al.22 resulted in a very different global P pattern: most widespread large deficits were in South America (North-Eastern countries, e.g., Argentina and Paraguay), the northern United States and Eastern Europe while the largest surpluses covered most of East Asia, Western and Southern Europe, the coastal United States, South-Eastern Brazil and Uruguay.
With average soil depletion due to erosion of 9.6 kg ha−1 yr−1 in Africa, the overall P balance is already negative by 9.7 kg ha−1 yr−1 today (Table 1, Fig. 4). As the average P depletion in Africa due to negative fluxes in organic P management equals the input fluxes from the atmosphere plus chemical fertilizer, African farmers could decrease P losses to near zero with effective soil erosion mitigation. Even though the system’s P depletion due to organic P management is relatively low in Africa (−2.2 kg ha−1 yr−1) compared to a global average (−5.2 kg ha−1 yr−1), the high overall P losses are unlikely to be covered neither from a mitigated and more sustainable organic P management nor from increased chemical fertilizer input. P fluxes due to organic P management are calculated here as the sum of manure and residue input minus plant uptake (which results in biomass export in arable systems with the exception of residues left on the field). The overall sum of plant uptake is likely to increase with increased need for food and feed parallel to a predicted population and livestock growth in Africa in the future. Many soils in sub-Saharan Africa have already been characterized as deficient for levels of plant-available P for the last decades53. Manure and residue input is simultaneously in demand in Africa today (shortage of biomass in general, low animal production and even if there is manure available, there are no means to transport it to where it is needed), which results in the recommendations of an integrated farm management with combinations of organic and inorganic fertilizers54,55,56. With the inorganic P fertilizers becoming increasingly scarce, the depletion due to organic P management can be expected to increase in Africa in the future. Simultaneously, today’s prices for chemical fertilizer can already be 2–6 times more expensive for a farmer in Africa than in Europe due to higher transport and storage costs3, even though Africa itself has the highest geological P deposits in the world (according to today’s estimates 80% of the global geological P deposits are located in Morocco and the Western Sarah8). As such, and if the political situation does not change dramatically (e.g., that the P supplies are marketed within Africa instead of being exported to US, Europe and China), the only realistic means of reducing P depletion of African soils today, and in the future, is to drastically reduce soil erosion.
We recognize that the values calculated in Table 1 and Fig. 4 are gross estimates over large scales and that spatial context and scale, especially on the African continent is important. P deficiency is a country, district, farm and soil specific issue in Africa, for example parts of east Africa and the Sahel have substantial deficiencies57. In sub Saharan Africa ~40% of soils are considered to have low nutrient reserves ( More

Regionally-downscaled climate change projections
We used a management strategy evaluation (MSE) applied to ensemble projections of a climate-enhanced multispecies stock assessment within the integrated modeling framework of the Alaska Climate Change Integrated Modeling project (ACLIM)19. For this, six high resolution downscaled projections of oceanographic and lower trophic level conditions in the Bering Sea (using the Regional Ocean Modeling System49,50) were coupled to the BESTNPZ nutrient-phytoplankton-zooplankton model51; we refer to this model complex throughout this paper as the Bering10K ROMSNPZ, or just ROMSNPZ, model. Boundary conditions were driven by three global general circulation models (GFDL-ESM2M52, CESM153, and MIROC-ESM54) projected (2006–2099) under the high-baseline emission scenario Representative Concentration Pathway 8.5 (RCP 8.5) and midrange global carbon mitigation (RCP 4.5; note, that for CESM under RCP 4.5, projections from 2080–2100 were unavailable so conditions from 2080–2099 were held constant at 2080 conditions for that scenario only) future scenarios from the Coupled Model Intercomparison Project phase 5 (CMIP5)29,55. Hermann et al.30,56 also report on downscaled hindcasts of oceanographic and lower trophic conditions in the EBS from 1970–2012 (see refs. 30,56,57 for detailed descriptions of model evaluation and performance). For each downscaled model simulation, we replicated the National Marine Fisheries Service Alaska Fisheries Science Center annual summer bottom-trawl survey in time and space in the ROMSNPZ model (using historical mean survey date at each latitude and longitude of each gridded survey station) to derive estimates of sea surface and bottom temperatures (Fig. 1). We additionally used a polygon mask of the survey area to estimate the average zooplankton abundance in the system during spring, summer, winter, and fall months. These indices were derived for each climate projection scenario, as well as a persistence scenario where conditions were held constant at the average of those for 2006–2017 from a hindcast simulation. All index projections were bias corrected to the 2006–2017 hindcast period using the delta method assuming unequal variance in the GCM projections and hindcast58 such that:

$$T_{mathop {{{mathrm{fut}}}}limits^prime ,y} = bar T_{{mathrm{hind}},overrightarrow {scriptstyle{mathrm{ref}}} } + frac{{sigma _{{mathrm{hind}},overrightarrow {scriptstyle{mathrm{ref}}} }}}{{sigma _{{mathrm{fut}}, overrightarrow{scriptstyle{mathrm{ ref}}} }}}left( {T_{{mathrm{fut}},y} – bar T_{{mathrm{fut}},overrightarrow {scriptstyle{mathrm{ref}}} }} right)$$
(1)

where (T_{mathop {{{mathrm{fut}}}}limits^prime ,y}) is the bias-corrected projected timeseries, (T_{{mathrm{fut}},y}) is the raw projected timeseries, (bar T_{{mathrm{hind}},overrightarrow {scriptstyle{mathrm{ref}}} }) is the mean of the hindcast during the reference years (overrightarrow {{mathrm{ref}}}) (2006–2017), (bar T_{{mathrm{fut}},overrightarrow {scriptstyle{mathrm{ref}}} }) is the mean of the raw projected timeseries during the reference years (overrightarrow {{mathrm{ref}}}), (sigma _{{mathrm{hind}},overrightarrow {scriptstyle{mathrm{ref}}} }) is the standard deviation of the hindcast during the reference years (overrightarrow {{mathrm{ref}}}), (sigma _{{mathrm{fut}},overrightarrow {scriptstyle{mathrm{ref}}} }) is the standard deviation of the raw projection timeseries during the reference years (overrightarrow {{mathrm{ref}}}).
Climate-enhanced multispecies stock assessment model
Bias-corrected indices were then used as covariates in the climate-enhanced multispecies stock assessment model for the Bering Sea (hereafter CEATTLE)59 to evaluate the performance of alternative management approaches on future fish biomass and catch. CEATTLE is a climate-enhanced multispecies statistical age-structured assessment model with parameters for growth that are functions of temperature (i.e., temperature-specific average weight-at-age) and predation that are functions of temperature (via a bioenergetics-based predation sub-model)59,60,61. Since 2016, the model has been used operationally in the Bering sea as a supplement to the annual BSAI pollock stock assessment61. Various configurations of CEATTLE are possible; for this study we chose one where temperature-specific predator and prey interactions influenced natural mortality, temperature influenced weight-at-age, and the spawner-recruit relationship was a function of physical and biological future conditions as well as random variability (i.e., a climate-informed multispecies model). We fit the model using penalized maximum likelihood to survey biomass, diet, and fishery harvest data for three groundfish species pollock, Pacific cod, and arrowtooth flounder from the EBS in the EBS over the period 1979–2017. We also used the Bering10K ROMSNPZ model to produce detailed hindcasts of temperature for the period 1970–2017. We used hindcast-extracted timeseries from the ROMSNPZ model and CEATTLE model estimates of recruitment ((R_{i,y,l})) and spawning biomass ((B_{i,y – 1})) in hindcast year y for each species i to fit a climate-enhanced logistic recruitment per spawner model36, such that:

$$ln left( {hat R_{i,y}} right) = alpha _i – beta _{0,i}B_{i,y – 1} + ln left( {B_{i,y – 1}} right) + {{{bf{B}}}}_i{{{bf{X}}}} + varepsilon _{i,y}$$
(2)

where ({{{bf{B}}}}_i{{{bf{X}}_l}}) is the summed product of each covariate parameter (beta _{ij}) and the corresponding environmental covariate (X_{j,y}) for each bias-corrected environmental index (j = ( {1,2…n_j} )). We selected indices representative of ecological conditions important for groundfish recruitment in the Bering sea39; spring and fall large zooplankton abundances, survey replicated bottom temperature, and extent of the residual cold pool of extremely dense and cold sea water that persists across the EBS shelf following spring sea ice retreat. We assumed normally distributed (in log space) residual errors for each species ((varepsilon _{i,y} sim Nleft( {0,sigma _i^2} right))). The CEATTLE model was then projected forward where ROMSNPZ indices from individual projections drove growth, predation, and recruitment in each future simulation year36,62.
Evaluation of harvest management approaches
Previous authors have defined EM (i.e., the incorporation of ecosystem information into marine resource management) as a continuum between two paradigms of management and focus18. On one end is within-sector single-species management that considers ecosystem information (EAFM) and on the other is cross-sectoral whole of ecosystem management (i.e., EBM). EBFM is intermediate between these and is defined by quantitative incorporation of ecosystem interactions into assessment models and target setting (EBFM). Most fisheries management in the Bering Sea can be characterized as EBFM or EAFM, with increasing trends toward cross-sectoral coordination at the scale of EBM. Here we focus on one aspect on this scale of potential management options, operational EBFM and EAFM as captured through the CEATTLE multispecies stock assessment model and harvest policies decisions made annually under the constraint of the 2 MT cap (modeled via the ATTACH model).
MSE is a process of “assessing the consequences of a range of management strategies or options and presenting the results in a way which lays bare the tradeoffs in performance across a range of management objectives”63. MSE has been frequently used to evaluate alternative management strategies based on single-species estimation methods64. It is increasingly used to evaluate ecosystem management performance, although these evaluations are far less commonplace due to the complexity of modeling and assessing the performance of ecosystem level metrics64. Importantly, MSE “does not seek to proscribe an optimal strategy or decision”63, rather it aims to describe the uncertainty and tradeoffs inherent in alternative strategies and scenarios. In this case, through a series of workshops, we worked with managers and stakeholders to identify priority scenarios and outputs19. From this, risk, sensitivity, and uncertainty under contrasting climate scenarios were requested outputs of the analysis, as was the performance of current climate-naive EBFM policies.
A key component of MSE is identifying and quantifying uncertainty (i.e., process, observation, estimation, model, and implementation error) and representing it using an operating model. In the case of this MSE, the focus was on process error uncertainty due to variation in recruitment about the fitted stock-recruitment relationship, one major source of model error in the form of alternative climate scenarios, and implementation error. The MSE does not account for estimation error (uncertainty in the parameters of the operating model) nor observation error. This is because the estimates of recruitment and spawning biomass from CEATTLE for the BSAI are very precise (see Fig. 10 in ref. 60) and the estimation and operating models are therefore very similar. Thus, CEATTLE is the operating model for this MSE and implicitly the estimation method. In this approach we assume that while allowing for observation error would have increased overall error, the effect would have been minor compared to the investigated uncertainties. Future analyses using a full MSE (i.e., separate operating and estimation models) could evaluate the effect of observation error, but perhaps more importantly, the potential for model error, whereby the population dynamics model (on which the estimation method is based) differs from that of the operating model such that the estimates on which management decisions are made are biased relative to the true values in the operating model.
Given this we summarized the relative change in catch and biomass for the three species in the model under the following fishing scenarios (Fig. 1): (a) projections without harvest ((F_{i,y} = 0)) in each year y of scenario l for each species i, (b) projections under target harvest rate (Supplementary Fig. 7 left) and with a sloping harvest control rule (HCR) (Supplementary Fig. 7 right), (c) as in 2 but with the constraint of a 2 MT cap applied dynamically to the three focal species only.
Under the North Pacific Fishery Management Council (NPFMC) constraint of the 2 MT cap on cumulative total annual catch, realized harvest (i.e., catch) and specification of individual species harvest limits known as Total Allowable Catch (TAC; metric tons) are a function of the acceptable biological catch (ABC) for the given species, as well as ABC of other valuable species in the aggregate complex19,34 (https://github.com/amandafaig/catchfunction). TAC must be set at or below ABC for each species, therefore TAC of individual species are traded-off with one another to avoid exceeding the 2 MT cap. From 1981 to 1983, the TAC of pollock was reduced significantly below the ABC and in 1984 the 2 MT cap became part of the BSAI fishery management plan21,34,65. Pacific cod regulations have changed markedly over recent decades and it was only in the 1990s that in many years the catch and TAC approached its ABC. Thus, we used the socioeconomic ATTACH model (the R package ATTACHv1.6.0 is available with permission at https://github.com/amandafaig/catchfunction34) to model realized catch in each simulation year as a function of CEATTLE assessment estimates of ABC (tons) for pollock, Pacific cod, and arrowtooth flounder under future projections (2018–2100). This entailed three steps for each future simulation year (y):
1.
project the population forward from (y – 1) to (y) using estimated parameters from the multispecies mode of the CEATTLE model fit to data from 1979 to 2017 and recruitment based on biomass in simulation year (y) and future environmental covariates from the ROMSNPZ model downscaled projections (see “Methods” above) to determine ({mathrm{ABC}}_{i,y,l}) for each species (i) under each scenario (l) given the sloping harvest control rule for pollock, Pacific cod, and arrowtooth flounder in each simulation year (y);

2.
use ({mathrm{ABC}}_{i,y,l}) of each species from step 1 as inputs to the ATTACH model in order to determine the North Pacific Marine Fishery Council Total Allowable Catch (TACi,y,l) for the given simulation l year y;

3.
use TACi,y,l from step 2 to estimate catch (tons) in the simulation year (Fig. 1); remove catch from the population and advance the simulation forward 1 year.

Determine the annual ABC
We used end-of-century projections (2095–2099) to derive a maximum sustainable yield (MSY) proxy for future harvest recommendations (ABCi,y,l) for each scenario l. To replicate current management, we used a climate-specific harvest control rule that uses climate-naive unfished and target spawning biomass reference points ((B_{0,i}) and (B_{{mathrm{target},i}}), respectively) and corresponding harvest rates ((F_{i,y} = 0) and (F_{i,y} = F_{{mathrm{target}}}) and (B_{i,y,l}) in each simulation l year (y) for each species (i)

$${mathrm{ABC}}_{i,y,l} = mathop {sum }limits_a^{A_i} left( {frac{{S_{i,a}F_{{mathrm{ABC}},i,y,l}}}{{Z_{i,a,y,l}}}left( {1 – e^{ – Z_{i,a,y,l}}} right)N_{i,a,y,l}W_{i,a,y,l}} right)$$
(3)

where (W_{i,a,y,l}), (N_{i,a,y}), and (Z_{i,a,y,l}) is the climate-simulation specific annual weight, number, and mortality (i.e., influenced through temperature effects on recruitment, predation, and growth) at age (a) for (A_i) ages in the model, and (S_{i,a}) is the average fishery age selectivity from the estimation period 1979–201759,60. (F_{{mathrm{ABC}},i,y,l}) and was determined in each simulation timestep using an iterative approach66 whereby we: (i) first determined average (B_{0,i}) values in years 2095–2099 by projecting the model forward without harvest (i.e., (F_{i,y} = 0)) for each species under the persistence scenario. We then (ii) iteratively solved for the harvest rate that results in an average spawning biomass ((B_{i,y})) during 2095–2099, that is, 40% of (B_{0,i}) (i.e., (F_{{mathrm{target}},i})) for pollock and Pacific cod simultaneously, with arrowtooth flounder (F_{i,y}) set to the historical average (as historical F for arrowtooth flounder ≪(F_{40% })); once (F_{{mathrm{target}},i}) for pollock and Pacific cod were found, we then iteratively solved for (F_{{mathrm{target}},i}) for arrowtooth flounder (Supplementary Fig. 7 left panel)59,60. Last, (iii) to derive a climate-informed ({mathrm{ABC}}_{i,y,l}) in each simulation year, the North Pacific Marine Fisheries Council (hereafter, “Council”) Tier 3 sloping harvest control rule with an ecosystem cutoff at (B_{20% }) was applied to adjust (F_{{mathrm{ABC}},i,y,l}) lower than (F_{{mathrm{target}},i}) if the simulation specific (climate-informed) (B_{i,y,l}) was lower than 40% of the climate-naive (B_{0,i}) at the start of a given year or set to 0 if (B_{i,y,l} , More

1.
IPCC. 2013 Supplement to the 2006 IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, Wetlands, 2014).
2.
Smith, P. et al. In Climate Change 2014: Mitigation of Climate Change. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds. Edenhofer, O. et al.) (Cambridge University Press, Cambridge, UK, 2014).

3.
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 (IPCC, Geneva, Switzerland, 2014).

4.
Ravishankara, A. R., Daniel, J. S. & Portmann, R. W. Nitrous oxide (N2O): the dominant ozone-depleting substance emitted in the 21st century. Science 326, 123–125 (2009).
CAS  Article  Google Scholar 

5.
Baumert, K., Herzog, T. & Pershing, J. Navigating the Numbers: Greenhouse Gas Data and International Climate Policy (World Resources Institute, Washington, DC, 2005).

6.
Oktarita, S., Hergoualch, K., Anwar, S. & Verchot, L. V. Substantial N2O emissions from peat decomposition and N fertilization in an oil palm plantation exacerbated by hotspots. Environ. Res. Lett. 12, 104007 (2017).
Article  Google Scholar 

7.
Yu, Z., Loisel, J., Brosseau, D. P., Beilman, D. W. & Hunt, S. J. Global peatland dynamics since the last glacial maximum. Geophys. Res. Lett. 37, L13402 (2010).
Google Scholar 

8.
Leifeld, J. & Menichetti, L. The underappreciated potential of peatlands in global climate change mitigation strategies. Nat. Commun. 9, 1071 (2018).
CAS  Article  Google Scholar 

9.
Limpens, J., Heijmans, M. M. P. D. & Berendse, F. In Boreal Peatland Ecosystems (eds Wieder, R. K. & Vitt, D. H.) 195 (Springer, Berlin, 2006).

10.
Joosten, H., Tapio-Biström, M. L. & Tol, S. Peatlands—Guidance for Climate Change Mitigation Through Conservation, Rehabilitation and Sustainable use 2nd edn (FAO & Wetlands International Landscape, ICIMOD, Kathmandu, 2012).

11.
Green, S. M. & Page, S. Tropical peatlands: current plight and the need for responsible management. Geol. Today 33, 174–179 (2017).
Article  Google Scholar 

12.
Moore, T. R. & Clarkson, B. R. Dissolved organic carbon in New Zealand peatlands. NZ J. Mar. Freshw. Res. 41, 137–141 (2007).

13.
Renou-Wilson, F., Barry, C., Müller, C. & Wilson, D. The impacts of drainage, nutrient status and management practice on the full carbon balance of grasslands on organic soils in a maritime temperate zone. Biogeosciences 11, 4361–4379 (2014).
Article  Google Scholar 

14.
Martikainen, P. J., Nykänen, H., Crill, P. & Silvola, J. Effect of a lowered water table on nitrous oxide fluxes from northern peatlands. Nature 366, 51–53 (1993).
CAS  Article  Google Scholar 

15.
Wrage-Mönnig, N. et al. Role of nitrifier denitrification in the production of nitrous oxide revisited. Soil Biol. Biochem. 123, A3–A16 (2018).
Article  Google Scholar 

16.
Pärn, J. et al. Nitrogen-rich organic soils under warm well-drained conditions are global nitrous oxide emission hotspots. Nat. Commun. 9, 1135 (2018).
Article  Google Scholar 

17.
Repo, M. E. et al. Large N2O emissions from cryoturbated peat soil in tundra. Nat. Geosci. 2, 189–192 (2009).
CAS  Article  Google Scholar 

18.
Klemedtsson, L., Von Arnold, K., Weslien, P. & Gundersen, P. Soil C/N ratio as a scalar parameter to predict nitrous oxide emissions. Global Change Biol. 11, 1142–1147 (2005).
Article  Google Scholar 

19.
Leppelt, T. et al. Nitrous oxide emission budgets and land-use-driven hotspots for organic soils in Europe. Biogeosciences 11, 6595–6612 (2014).
Article  Google Scholar 

20.
Petersen, S. O. et al. Annual emissions of CH4 and N2O, and ecosystem respiration, from eight organic soils in western Denmark managed by agriculture. Biogeosciences 9, 403–422 (2012).
CAS  Article  Google Scholar 

21.
Leifeld, J. Distribution of nitrous oxide emissions from managed organic soils under different land uses estimated by the peat C/N ratio to improve national GHG inventories. Sci. Total Environ. 631–632, 23–26 (2018).
Article  Google Scholar 

22.
van Beek, C. L. et al. Emissions of N2O from fertilized and grazed grassland on organic soil in relation to groundwater level. Nutr. Cycling Agroecosyst. 86, 331–340 (2010).
Article  Google Scholar 

23.
Maljanen, M. et al. Afforestation does not necessarily reduce nitrous oxide emissions from managed boreal peat soils. Biogeochemistry 108, 199–218 (2012).
CAS  Article  Google Scholar 

24.
Liimatainen, M. et al. Factors controlling nitrous oxide emissions from managed northern peat soils with low carbon to nitrogen ratio. Soil Biol. Biochem. 122, 186–195 (2018).
CAS  Article  Google Scholar 

25.
Säurich, A., Tiemeyer, B., Dettmann, U. & Don, A. How do sand addition, soil moisture and nutrient status influence greenhouse gas fluxes from drained organic soils? Soil Biol. Biochem. 135, 71–84 (2019).
Article  Google Scholar 

26.
Laine, J. et al. Effect of water-level drawdown on global climatic warming: northern peatlands. Ambio 25, 179–184 (1996).
Google Scholar 

27.
Laudone, G. M. et al. A model to predict the effects of soil structure on denitrification and N2O emission. J. Hydrol. 409, 283–290 (2011).
CAS  Article  Google Scholar 

28.
Wu, L. et al. Simulation of nitrous oxide emissions at field scale using the SPACSYS model. Sci. Total Environ. 530–531, 76–86 (2015).
Article  Google Scholar 

29.
Couwenberg, J. et al. Assessing greenhouse gas emissions from peatlands using vegetation as a proxy. Hydrobiologia 674, 67–89 (2011).
CAS  Article  Google Scholar 

30.
Liu, H., Zak, D., Rezanezhad, F. & Lennartz, B. Soil degradation determines release of nitrous oxide and dissolved organic carbon from peatlands. Environ. Res. Lett. 14, 094009 (2019).
CAS  Article  Google Scholar 

31.
Tanneberger, F., Joosten, H., Moen, A. & Whinam, J. In Mires and Peatlands of Europe—Status, Distribution and Conservation (eds Joosten, H., Tanneberger, F. & Moen, A.) 173–196 (Schweizerbart Science Publishers, Stuttgart, 2017).

32.
Renou-Wilson, F. et al. Rewetting degraded peatlands for climate and biodiversity benefits: results from two raised bogs. Ecol. Eng. 127, 547–560 (2019).
Article  Google Scholar 

33.
Lamers, L. P. M. et al. Ecological restoration of rich fens in Europe and North America: From trial and error to an evidence-based approach. Biol. Rev. Camb. Philos. Soc. 90, 182–203 (2015).
Article  Google Scholar 

34.
Tiemeyer et al. A new methodology for organic soils in national greenhouse gas inventories: Data synthesis, derivation and application. Ecol. Indic. 109, 105838 (2020).
CAS  Article  Google Scholar 

35.
Wilson et al. Multiyear greenhouse gas balances at a rewetted temperate peatland. Glob. Change Biol. 22, 4080–4095 (2016).
Article  Google Scholar 

36.
Evans, C. et al. Implementation of an Emission Inventory for UK Peatlands. Report to the Department for Business, Energy and Industrial Strategy, Centre for Ecology and Hydrology, Bangor. 88 (2017).

37.
Günther, A. et al. Prompt rewetting of drained peatlands reduces climate warming despite methane emissions. Nat. Commun. 11, 1644 (2020).
Article  Google Scholar 

38.
Nykänen, H., Alm, J., Lång, K., Silvola, J. & Martikainen, P. J. Emissions of CH4, N2O and CO2 from a virgin fen and a fen drained for grassland in Finland. J. Biogeogr. 22, 351–357 (1995).
Article  Google Scholar 

39.
Drösler, M. et al. Klimaschutz furch Moorschutz in der Praxis (Thünen-Institut fur Agrarklimaschutz, Brauschweig, Germany, 2013).

40.
Mojeremane, W., Rees, R. M. & Mencuccini, M. The effects of site preparation practices on carbon dioxide, methane and nitrous oxide fluxes from a peaty gley soil. Forestry 85, 1–15 (2012).
Article  Google Scholar 

41.
Pronger, J., Schipper, L. A., Hill, R. B., Campbell, D. I. & McLeod, M. Subsidence rates of drained agricultural peatlands in New Zealand and the relationship with time since drainage. J. Environ. Qual. 43, 1442 (2014).
CAS  Article  Google Scholar 

42.
Hume, N. P., Fleming, M. S. & Horne, A. J. Plant carbohydrate limitation on nitrate reduction in wetland microcosms. Water Res. 36, 577–584 (2002).
CAS  Article  Google Scholar 

43.
Höper, H. et al. In Peatlands and Climate Change (ed. Strack, M.) 182–210 (International Peat Society, Jyväskylä, Finland, 2008).

44.
Davidson, E. A. The contribution of manure and fertilizer nitrogen to atmospheric nitrous oxide since 1860. Nat. Geosci. 2, 659–662 (2009).
CAS  Article  Google Scholar 

45.
Tiemeyer, B. et al. High emissions of greenhouse gases from grasslands on peat and other organic soils. Global Change Biol. 22, 4134–4149 (2016).
Article  Google Scholar 

46.
Andersen, R. et al. An overview of the progress and challenges of peatland restoration in Western Europe. Restor. Ecol. 25, 271–282 (2016).
Article  Google Scholar 

47.
Lugato, E., Paniagua, L., Jones, A., de Vries, W. & Leip, A. Complementing the topsoil information of the Land Use/Land Cover Area Frame Survey (LUCAS) with modelled N2O emissions. PLoS ONE 12, e0176111 (2017).
Article  Google Scholar 

48.
Xu, J., Morrisa, P. J., Liu, J. & Holden, J. PEATMAP: Refining estimates of global peatland distribution based on a meta-analysis. Catena 160, 134–140 (2018).
Article  Google Scholar 

49.
Pflugmacher, D., Rabe, A., Peters, M. & Hostert, P. Mapping pan-European land cover using Landsat spectral-temporal metrics and the European LUCAS survey. Remote Sens. Environ. 221, 583–595 (2019).
Article  Google Scholar 

50.
Hierderer, R. EFSA Spatial Data Version 1.1, Data Properties and Processing (Publication Office of the European Union, Luxembourg, 2012).

51.
Jones, R. J., Hiederer, R., Rusco, E. & Montanarella, L. Estimating organic carbon in the soils of Europe for policy support. Eur. J. Soil Sci. 56, 655–671 (2005).
CAS  Article  Google Scholar 

52.
Joosten, H., Tannenberger, F. & Moen, A. Mires and Peatlands of Europe (Schweizerbart Science Publishers, Stuttgart, Germany, 2017).

53.
Liu, H., Price, J. S., Rezanezhad, F. & Lennartz, B. Century-scale shifts in peat hydro-physical properties as induced by drainage Water Resource Research (2020).

54.
Figueres, C. et al. Three years to safeguard our climate. Nature 546, 593–595 (2017).
CAS  Article  Google Scholar  More

Trabaque tufa record
Trabaque Canyon (40.36° N; 2.26° W; 840 m above sea level) is located in the central Iberian Peninsula (Fig. 1). At this site, tufa deposits precipitate as freshwater carbonates downstream of overflow karst springs. During the last interglacial period, tufa precipitated continuously at the studied site while water level of the aquifer was high enough for upstream springs to discharge13. Outcrops of the studied tufa deposit are preserved in the margins of Trabaque River over a distance of 500 m downstream of overflow karstic springs. The studied tufa deposit is 12 m thick, with a gentle ramp morphology, and a simple stratigraphy of sub-horizontal tufa beds that covered the full section of the narrow canyon. The accumulation of tufa created a small lake upstream the ramp, which prevented erosive events while the deposit was active, because most of the river bedload was accumulated in the basin of the lake. This configuration favoured the lack of erosive episodes in the tufa and the deposition of a continuous record. The tufa deposit was partially eroded by subsequent fluvial incision once the tufa accretion ceased and detrital sediments filled the lake basin and started to flow over the ramp during floods. The tufa deposit is mostly composed of well-cemented intra-clastic and peloidal carbonate particles13. The deposit comprises tufa beds 0.02–1 m thick that typically extend tens of metres downstream. At the base of the section, the tufa lies over loose fluvial sediments of sandy silt, whereas at the top of the section there is an erosive scar, and recent gravitational deposits overlay the tufa preserved in the slopes of the canyon.
Figure 1

Pictures of Trabaque Canyon and the studied deposit. (a) Trabaque Canyon. The river flows according to yellow arrows. The red ellipse shows the location of the main section where the deposit was sampled. The inlet map shows the location of Trabaque Canyon within the Iberian Peninsula. (b) View of most of the studied Trabaque tufa section. The base and top of the section are missing from this panorama. The centre of the valley bottom is to the left of the image and the slope of the canyon to the right. The river flowed from the position of the observer towards the tufa deposit. The picture shows gravitational pulses GP-2 and GP-3 that interdigitate with the tufa deposit, and their disappearance from the bottom of the valley after GP-3. (c) Detail of GP-3 gravitational deposit. (d) Detail of the alternation between well-cemented and loose tufa beds at the top of the section.

Full size image

The base of the deposit section is characterized by nearly 4 m of tufa sediments in the centre of the valley, laterally interdigitating with gravitational deposits towards the slopes (Fig. 1). These gravitational deposits partially invaded the bottom of the valley during three distinct pulses. These gravitational deposits occurred during periods of enhanced slope processes due to the decrease in vegetation cover on the canyon slopes during prolonged dry periods. The evidence of local erosion recorded by the gravitational deposits is consistent with other proxies that record local and regional erosion and that are displayed in Fig. 2. Thus, independent evidence of erosion is also recognized from the increase of insoluble residue (IR) particles in the tufa, recorded by the percentage of silt IR. IR particles were transported to the tufa by the river or by the action of wind. The increase of these particles in the tufa is interpreted as enhanced erosion, not only from the catchment but also from outside the basin. Higher concentrations of Si and Al are also interpreted as proxies of soil erosion from areas with silicate substrates inside or outside the catchment. The increase of micro-charcoal particles in the tufa is also interpreted as a sign of enhanced soil erosion. Charcoals were incorporated to the tufa during floods or transported by the wind after the occurrence of fires, as well as from the erosion of soils that accumulated charcoals from previous fire events. In any case, the increase of micro-charcoals in the tufa record suggests soil erosion due to the lost of vegetation cover. Major events of local and regional erosion occurred synchronously (Fig. 2), supporting that the common decreases in vegetation cover that resulted in erosion events were related to periods of reduced precipitation.
Figure 2

Record of the Trabaque tufa deposit. (a) Simplified lithological log of the Trabaque record. Patterns represent gravitational deposits (black) with distinct three pulses, well-cemented tufa sediments (light grey), and alternation of well-cemented and loose tufa sediments (dark grey). (b,c) Tufa δ18O and δ13C records. Isotope values at each date (dots) are the average of 3 sub-samples and blue/red line is a 3-point running mean. The grey shade shows the 1σ variability of the three sub-samples along the record. (d) Concentration of Si and Al. (e) Silt-sized insoluble residue in tufa as percentage of the total sample. (f) Counts of micro-charcoal particles  More

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript. More

General situation of the research area
The research area was conducted at Liufangzi village, Gongzhuling city, Jilin Province (N43°34′10″, E124°52′55″), as shown in Fig. 8. The area has a continental monsoon climate in the humid area of the middle temperate zone, with an average annual precipitation of 594.8 mm, which is mainly concentrated in June and August. The average annual temperature is 5.6 °C, and the daily average temperature drops to 0 °C in November of each year, with a freezing period of up to five months. Corn is one of the main commodity crops in the area, with a sowing date in early May and a harvest date in early October.
Figure 8

Location of study area (Liufangzi Village, Gongzhuling City, Jilin Province).

Full size image

The soil of the site is a silty loam black soil, which had been planted with monoculture corn with no tillage for 5 years. On October 5, 2018 (after the autumn harvest), a flat field was selected to set up the experiment. Soil samples were collected using the zigzag sampling method, and selective physical and chemical properties of soil were determined, including pH (5.48), organic matter (26.4 g kg−1), clay (29.12%), and soil bulk density (1.21 g cm−3 in 5–10 cm and 1.53 g cm−3 in 20–25 cm).
Reagents and instruments
Reagents
The main raw material of the added impervious agent was corn starch and acrylic compound, which was entrusted to Jilin Yida Chemical Co., Ltd. The added urea was an analytical reagent, and the reagents used for analysis included H2SO4, H3PO4, NaOH, NH4OH, NH4Cl, K2S2O8, Na2B4O7, KNO3, KNO2, K2Cr2O7, FeSO4, sulfonamide, and naphthalene ethylenediamine hydrochloride; these were all analytical reagents provided by Beijing Chemical Plant.
Instruments laboratory-built soil leaching column; continuous flow injection analyser (SKALAR SA++, Netherlands).
Test plot setup and agronomic practices
The experimental plots were maintained in the field consisting of (1) CK (no-tillage control treatment, with corn straw removed and soil left under no-till management); (2) ploughing treatment (corn straw was removed and then mouldboard ploughed to a 30 cm depth); (3) straw returning treatment (corn straw (25.32% moist) was incorporated into the soil on October 5, 2018 (after autumn harvest), with an application amount of 1.25 kg m−2. Briefly, corn straw was chopped into small pieces (0.5 cm length), evenly placed on the soil surface, and then incorporated into the soil with ploughing (the depth of 30 cm)); and (4) impervious agent addition treatment (the impervious agent mentioned previously evenly laid on the soil surface at the amount of 15 g m−2 and then incorporated into the 0–30 cm soil by mouldboard ploughing). The abovementioned field operations were conducted after corn harvest in the fall of 2018 with a testing area of 10 m × 50 m for each plot and three replicates for each treatment. In the following spring (2019), grain corn was planted in all treatment plots with a planting density of 65,000 plants ha−1. All plots were managed in the same way with a one-time fertilization application of 200–90-90 kg (N-P-K) ha−1 and 2,4-d spray as weed control.
For all the above treatments (including the control treatment), undisturbed soils (0–30 cm layer) were collected with an undisturbed soil column (refer to Fig. 9) for the leaching experiment on September 25, 2019 (before autumn harvest, after 350 days of straw returning to the field); soil samples of 0–15 cm were collected for determination of soil organic matter and adsorption experiment of nitrogen in the soil; and soil samples of 5–10 cm and 20–25 cm layers were collected for determination of soil bulk density. In addition, for the straw returning treatment, one sampling was added on May 25, 2019 (one month after sowing, 230 days after straw returning), for the determination of soil organic matter content and soil bulk density, nitrogen adsorption and leaching experiment in soil.
Figure 9

Schematic diagram of simulated leaching device of undisturbed soil column. (a) Soil extraction; (b) leaching; (c) physical map of leaching in undisturbed soil column. 1: Handle; 2.3.4: guide port; 5.6: screw port; 7: punching plate. I main body of leaching column; II soil cutter; III leaching solution collector.

Full size image

The soil samples used for soil organic matter determination and nitrogen absorption testing were air dried, sieved through a 2-mm sieve and visible plant debris and stones were removed, and then stored.
Experiment of nitrogen adsorption in soil
Ten parts of the soil samples (air-dried,  More

1.
Erős, T., O’Hanley, J. R. & Czeglédi, I. A unified model for optimizing riverscape conservation. J. Appl. Ecol. 55, 1871–1883 (2018).
Google Scholar 
2.
Ruggeri, P., Pasternak, E. & Okamura, B. To remain or leave: Dispersal variation and its genetic consequences in benthic freshwater invertebrates. Ecol. Evol. 9, 12069–12088 (2019).
PubMed  PubMed Central  Google Scholar 

3.
Baguette, M., Blanchet, S., Legrand, D., Stevens, V. M. & Turlure, C. Individual dispersal, landscape connectivity and ecological networks. Biol. Rev. 88, 310–326 (2013).
PubMed  Google Scholar 

4.
Geist, J. Seven steps towards improving freshwater conservation. Aquat. Conserv. Mar. Freshw. Ecosyst. 25, 447–453 (2015).
Google Scholar 

5.
Kujala, H., Lahoz-Monfort, J. J., Elith, J. & Moilanen, A. Not all data are equal: Influence of data type and amount in spatial conservation prioritisation. Methods Ecol. Evol. 9, 2249–2261 (2018).
Google Scholar 

6.
Johnson, J. B., Peat, S. M. & Adams, B. J. Where’s the ecology in molecular ecology?. Oikos 118, 1601–1609 (2009).
Google Scholar 

7.
Janse, J. H. et al. GLOBIO-aquatic, a global model of human impact on the biodiversity of inland aquatic ecosystems. Environ. Sci. Policy 48, 99–114 (2015).
Google Scholar 

8.
Newbold, T. et al. Global effects of land use on local terrestrial biodiversity. Nature 520, 45–50 (2015).
ADS  PubMed  CAS  Google Scholar 

9.
Moore, D., Cranston, G., Reed, A. & Galli, A. Projecting future human demand on the Earth’s regenerative capacity. Ecol. Indic. 16, 3–10 (2012).
Google Scholar 

10.
Yawson, D. O., Adu, M. O. & Armah, F. A. Impacts of climate change and mitigation policies on malt barley supplies and associated virtual water flows in the UK. Sci. Rep. 10, 1–12 (2020).
Google Scholar 

11.
Naidoo, R. et al. Global mapping of ecosystem services and conservation priorities. Proc. Natl. Acad. Sci. USA 105, 9495–9500 (2008).
ADS  PubMed  CAS  Google Scholar 

12.
Hermoso, V., Villero, D., Clavero, M. & Brotons, L. Spatial prioritisation of EU’s LIFE-Nature programme to strengthen the conservation impact of Natura 2000. J. Appl. Ecol. 55, 1575–1582 (2018).
Google Scholar 

13.
Hermoso, V., Morán-Ordóñez, A., Canessa, S. & Brotons, L. Realising the potential of Natura 2000 to achieve EU conservation goals as 2020 approaches. Sci. Rep. 9, 1–10 (2019).
CAS  Google Scholar 

14.
Lobera, G., Pardo, I., García, L. & García, C. Disentangling spatio-temporal drivers influencing benthic communities in temporary streams. Aquat. Sci. 81, 1–17 (2019).
CAS  Google Scholar 

15.
Richman, N. I. et al. Multiple drivers of decline in the global status of freshwater crayfish (Decapoda: Astacidea). Philos. Trans. R. Soc. B Biol. Sci. 370, 20140060 (2015).

16.
Manenti, R. et al. Causes and consequences of crayfish extinction: Stream connectivity, habitat changes, alien species and ecosystem services. Freshw. Biol. 64, 284–293 (2019).
Google Scholar 

17.
Kozák, P., Füreder, L., Kouba, A., Reynolds, J. & Souty-Grosset, C. Current conservation strategies for European crayfish. Knowl. Manag. Aquat. Ecosyst. 01, https://doi.org/10.1051/kmae/2011018 (2011).

18.
Pârvulescu, L. Introducing a new Austropotamobius crayfish species (Crustacea, Decapoda, Astacidae): A miocene endemism of the Apuseni Mountains, Romania. Zool. Anz. 279, 94–102 (2019).
Google Scholar 

19.
Kouba, A., Petrusek, A. & Kozák, P. Continental-wide distribution of crayfish species in Europe: Update and maps. Knowl. Manag. Aquat. Ecosyst. 413, 05–31 (2014).
Google Scholar 

20.
Pârvulescu, L. et al. A journey on plate tectonics sheds light on European crayfish phylogeography. Ecol. Evol. 9, 1957–1971 (2019).
PubMed  PubMed Central  Google Scholar 

21.
Pârvulescu, L. & Zaharia, C. Current limitations of the stone crayfish distribution in Romania: Implications for its conservation status. Limnologica 43, 143–150 (2013).
Google Scholar 

22.
Klobučar, G. I. V. et al. Role of the Dinaric Karst (western Balkans) in shaping the phylogeographic structure of the threatened crayfish Austropotamobius torrentium. Freshw. Biol. 58, 1089–1105 (2013).
Google Scholar 

23.
Qian, S. S., Cuffney, T. F., Alameddine, I., McMahon, G. & Reckhow, K. H. On the application of multilevel modeling in environmental and ecological studies. Ecology 91, 355–361 (2010).
PubMed  Google Scholar 

24.
Manning, P. et al. Redefining ecosystem multifunctionality. Nat. Ecol. Evol. 2, 427–436 (2018).
PubMed  Google Scholar 

25.
Koizumi, I., Usio, N., Kawai, T., Azuma, N. & Masuda, R. Loss of genetic diversity means loss of geological information: The endangered Japanese crayfish exhibits remarkable historical footprints. PLoS ONE 7, e33986 (2012).
ADS  PubMed  PubMed Central  CAS  Google Scholar 

26.
McNyset, K. M. Use of ecological niche modelling to predict distributions of freshwater fish species in Kansas. Ecol. Freshw. Fish 14, 243–255 (2005).
Google Scholar 

27.
Henrys, P. A. & Jarvis, S. G. Integration of ground survey and remote sensing derived data: Producing robust indicators of habitat extent and condition. Ecol. Evol. 9, 8104–8112 (2019).
PubMed  PubMed Central  Google Scholar 

28.
Pârvulescu, L., Zaharia, C., Satmari, A. & Drăguţ, L. Is the distribution pattern of the stone crayfish in the Carpathians related to karstic refugia from Pleistocene glaciations?. Freshw. Sci. 32, 1410–1419 (2013).
Google Scholar 

29.
Longshaw, M. & Stebbing, P. Biology and Ecology of Crayfish. (CRC Press, 2015).

30.
Chucholl, C. The bad and the super-bad: Prioritising the threat of six invasive alien to three imperilled native crayfishes. Biol. Invasions 18, 1967–1988 (2016).
Google Scholar 

31.
Chucholl, C. & Schrimpf, A. The decline of endangered stone crayfish (Austropotamobius torrentium) in southern Germany is related to the spread of invasive alien species and land-use change. Aquat. Conserv. Mar. Freshw. Ecosyst. 26, 44–56 (2016).
Google Scholar 

32.
Pârvulescu, L. et al. Flash-flood potential: A proxy for crayfish habitat stability. Ecohydrology 9, 1507–1516 (2016).
Google Scholar 

33.
Farr, T. G. et al. The shuttle radar topography mission. Rev. Geophys. 45, RG2004 (2007).

34.
Şandric, I. et al. Integrating catchment land cover data to remotely assess freshwater quality: A step forward in heterogeneity analysis of river networks. Aquat. Sci. 81, 26 (2019).
Google Scholar 

35.
Burkhard, B., Kroll, F., Nedkov, S. & Müller, F. Mapping ecosystem service supply, demand and budgets. Ecol. Indic. 21, 17–29 (2012).
Google Scholar 

36.
Zeller, K. A., McGarigal, K. & Whiteley, A. R. Estimating landscape resistance to movement: A review. Landsc. Ecol. 27, 777–797 (2012).
Google Scholar 

37.
Liaw, A. & Wiener, M. Classification and regression by randomForest. R News 2, 18–22 (2002).
Google Scholar 

38.
R Core Team. R: A Language and Environment for Statistical Computing. (R Foundation for Statistical Computing, 2017).

39.
Freeman, E. A. & Moisen, G. G. A comparison of the performance of threshold criteria for binary classification in terms of predicted prevalence and kappa. Ecol. Modell. 217, 48–58 (2008).
Google Scholar 

40.
Iorgu, E. I., Popa, O. P., Petrescu, A.-M. & Popa, L. O. Cross-amplification of microsatellite loci in the endangered stone-crayfish Austropotamobius torrentium (Crustacea: Decapoda). Knowl. Manag. Aquat. Ecosyst. 08, https://doi.org/10.1051/kmae/2011021 (2011).

41.
Peakall, R. & Smouse, P. E. genalex 6: Genetic analysis in Excel. Population genetic software for teaching and research. Mol. Ecol. Notes 6, 288–295 (2006).

42.
Goudet, J. FSTAT (Version 1.2): A computer program to calculate F-statistics. J. Hered. 86, 485–486 (1995).

43.
Rousset, F. genepop’007: A complete re-implementation of the genepop software for Windows and Linux. Mol. Ecol. Resour. 8, 103–106 (2008).
PubMed  Google Scholar 

44.
Van Oosterhout, C., Hutchinson, W. F., Wills, D. P. & Shipley, P. micro-checker: Software for identifying and correcting genotyping errors in microsatellite data. Mol. Ecol. Notes 4, 535–538 (2004).
Google Scholar 

45.
Dempster, A. P., Laird, N. M. & Rubin, D. B. Maximum likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. Ser. B 39, 1–22 (1977).
MathSciNet  MATH  Google Scholar 

46.
Chapuis, M. P. & Estoup, A. Microsatellite null alleles and estimation of population differentiation. Mol. Biol. Evol. 24, 621–631 (2007).
PubMed  CAS  Google Scholar 

47.
Weir, B. S. & Cockerham, C. C. Estimating F‐statistics for the analysis of population structure. Evolution (N. Y). 38, 1358–1370 (1984).

48.
Hammer, D. A. T., Ryan, P. D., Hammer, Ø. & Harper, D. A. T. Past: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica vol. 4 https://palaeo-electronica.orghttp//palaeo-electronica.org/2001_1/past/issue1_01.htm. (2001).

49.
Nei, M., Tajima, F. & Tateno, Y. Accuracy of estimated phylogenetic trees from molecular data. J. Mol. Evol. 19, 153–170 (1983).
ADS  PubMed  CAS  Google Scholar 

50.
Langella, O. Populations, 1.2. 30. https://bioinformatics.org/~tryphon/populations (1999).

51.
Pritchard, J. K., Stephens, M., Rosenberg, N. A. & Donnelly, P. Association mapping in structured populations. Am. J. Hum. Genet. 67, 170–181 (2000).
PubMed  PubMed Central  CAS  Google Scholar 

52.
Evanno, G., Regnaut, S. & Goudet, J. Detecting the number of clusters of individuals using the software STRUCTURE: A simulation study. Mol. Ecol. 14, 2611–2620 (2005).
PubMed  CAS  Google Scholar 

53.
Kopelman, N. M., Mayzel, J., Jakobsson, M., Rosenberg, N. A. & Mayrose, I. Clumpak: A program for identifying clustering modes and packaging population structure inferences across K. Mol. Ecol. Resour. 15, 1179–1191 (2015).
PubMed  PubMed Central  CAS  Google Scholar 

54.
Vähä, J. P. & Primmer, C. R. Efficiency of model-based Bayesian methods for detecting hybrid individuals under different hybridization scenarios and with different numbers of loci. Mol. Ecol. 15, 63–72 (2005).
Google Scholar 

55.
Bergl, R. A. & Viglant, L. Genetic analysis reveals population structure and recent migration within the highly fragmented range of the Cross River gorilla (Gorilla gorilla diehli). Mol. Ecol. 16, 501–516 (2006).
Google Scholar 

56.
Jombart, T., Devillard, S. & Balloux, F. Discriminant analysis of principal components: A new method for the analysis of genetically structured populations. BMC Genet. 11, 1–15 (2010).
Google Scholar 

57.
Paetkau, D., Calvert, W., Stirling, I. & Strobeck, C. Microsatellite analysis of population structure in Canadian polar bears. Mol. Ecol. 4, 347–354 (1995).
PubMed  CAS  Google Scholar 

58.
Duchesne, P. & Turgeon, J. FLOCK Provides Reliable Solutions to the ‘“Number of Populations”’ Problem. https://doi.org/10.1093/jhered/ess038.

59.
Janes, J. K. et al. The K = 2 conundrum. Mol. Ecol. 26, 3594–3602 (2017).
PubMed  Google Scholar 

60.
Funk, S. M. et al. Major inconsistencies of inferred population genetic structure estimated in a large set of domestic horse breeds using microsatellites. Ecol. Evol. 10, 4261–4279 (2020).
PubMed  PubMed Central  Google Scholar 

61.
Berger, C., Štambuk, A., Maguire, I., Weiss, S. & Füreder, L. Integrating genetics and morphometrics in species conservation—A case study on the stone crayfish, Austropotamobius torrentium. Limnologica 69, 28–38 (2018).
Google Scholar 

62.
Iojă, C. I. et al. The efficacy of Romania’s protected areas network in conserving biodiversity. Biol. Conserv. 143, 2468–2476 (2010).
Google Scholar 

63.
Rabăgia, T. & Maţenco, L. Tertiary tectonic and sedimentological evolution of the South Carpathians foredeep: Tectonic vs eustatic control. Mar. Pet. Geol. 16, 719–740 (1999).

64.
Rãdoane, M., Rãdoane, N. & Dumitriu, D. Geomorphological evolution of longitudinal river profiles in the Carpathians. Geomorphology 50, 293–306 (2003).
ADS  Google Scholar 

65.
Helms, B., Loughman, Z. J., Brown, B. L. & Stoeckel, J. Recent advances in crayfish biology, ecology, and conservation. Freshw. Sci. 32, 1273–1275 (2013).
Google Scholar 

66.
Svobodová, J. et al. The relationship between water quality and indigenous and alien crayfish distribution in the Czech Republic: Patterns and conservation implications. Aquat. Conserv. Mar. Freshw. Ecosyst. 22, 776–786 (2012).
Google Scholar 

67.
Pöckl, M. & Streissl, F. Austropotamobius torrentium as an indicator for habitat quality in running waters? Bull. Français la Pêche la Piscic. 743–758, https://doi.org/10.1051/kmae:2005030 (2005).

68.
Magyar, I. et al. Progradation of the paleo-Danube shelf margin across the Pannonian Basin during the Late Miocene and Early Pliocene. Glob. Planet. Change 103, 168–173 (2013).
ADS  Google Scholar 

69.
Zhang, Y., Luan, P., Ren, G., Hu, G. & Yin, J. Estimating the inbreeding level and genetic relatedness in an isolated population of critically endangered Sichuan taimen (Hucho Bleekeri) using genome-wide SNP markers. Ecol. Evol. 10, 1390–1400 (2020).
PubMed  PubMed Central  Google Scholar 

70.
Hoarau, G. et al. Low effective population size and evidence for inbreeding in an overexploited flatfish, plaice (Pleuronectes platessa L.). Proc. Biol. Sci. 272, 497–503 (2005).

71.
Jourdan, J. et al. Reintroduction of freshwater macroinvertebrates: Challenges and opportunities. Biol. Rev. https://doi.org/10.1111/brv.12458 (2018).
Article  PubMed  Google Scholar 

72.
Oidtmann, B., Heitz, E., Rogers, D. & Hoffmann, R. Transmission of crayfish plague. Dis. Aquat. Organ. 52, 159–167 (2002).
PubMed  Google Scholar 

73.
Rusch, J. C. et al. Simultaneous detection of native and invasive crayfish and Aphanomyces astaci from environmental DNA samples in a wide range of habitats in Central Europe. NeoBiota (2020).

74.
Hall, Q. A., Curtis, J. M., Williams, J. & Stunz, G. W. The importance of newly-opened tidal inlets as spawning corridors for adult Red Drum (Sciaenops ocellatus). Fish. Res. 212, 48–55 (2019).
Google Scholar 

75.
Stewart, F. E. C., Darlington, S., Volpe, J. P., McAdie, M. & Fisher, J. T. Corridors best facilitate functional connectivity across a protected area network. Sci. Rep. 9, 10852 (2019).
ADS  PubMed  PubMed Central  Google Scholar 

76.
Strauss, A., White, A. & Boots, M. Invading with biological weapons: The importance of disease-mediated invasions. Funct. Ecol. 26, 1249–1261 (2012).
Google Scholar 

77.
Clavero, M. & García-Berthou, E. Invasive species are a leading cause of animal extinctions. Trends Ecol. Evol. 20, 110 (2005).
PubMed  Google Scholar 

78.
Nunes, A. L., Tricarico, E., Panov, V. E., Cardoso, A. C. & Katsanevakis, S. Pathways and gateways of freshwater invasions in Europe. Aquat. Invasions 10, 359–370 (2015).
Google Scholar 

79.
Zeng, Y. & Yeo, D. C. J. Assessing the aggregated risk of invasive crayfish and climate change to freshwater crabs: A Southeast Asian case study. Biol. Conserv. 223, 58–67 (2018).
Google Scholar 

80.
Alonso, F., Temino, C. & Diéguez-Uribeondo, J. Status of the white-clawed crayfish, Austropotamobius pallipes (Lereboullet, 1858), in Spain: Distribution and legislation. 31–53 (2000).

81.
Van Dyck, H. & Baguette, M. Dispersal behaviour in fragmented landscapes: Routine or special movements?. Basic Appl. Ecol. 6, 535–545 (2005).
Google Scholar 

82.
Rodrigues, A. S. L., Pilgrim, J. D., Lamoreux, J. F., Hoffmann, M. & Brooks, T. M. The value of the IUCN Red List for conservation. Trends Ecol. Evol. 21, 71–76 (2006).
PubMed  Google Scholar 

83.
Füreder, L., Gherardi, F. & Souty-Grosset, C. Austropotamobius torrentium. The IUCN Red List of Threatened Species 2010 e.T2431A9439449 https://doi.org/10.2305/IUCN.UK.2010-3.RLTS.T2431A9439449.en (2010). More