Ecology

Quantifying heating of coastal ecosystemsA variety of metrics have been developed to quantify MHWs, but to date, they have mainly focused on surface heating evident from SST. Surface MHWs have been defined based on exceedance of 90% confidence intervals calculated seasonally from historical SST10,11. Defined in this way, the threshold temperature for quantifying a MHW is likely to vary seasonally from, for example, winter to summer. However, the physiological and ecological relevance of seasonally changing thresholds remains unclear, especially for tropical biota such as corals that typically inhabit a relatively narrow range of temperatures close to their physiological thermal limits12,13,14. Assessing heating in the specific context of temperature-induced coral bleaching has instead focused on calculating cumulative degree heating, sensitive to both the magnitude and duration of heating, above a fixed, putative ‘bleaching threshold’ defined by the local Maximum Monthly Mean SST (MMM); i.e., the mean summer-time peak temperature predicted to initiate coral stress and bleaching. Such heat accumulation has most commonly been expressed as Degree Heating Weeks (DHW in °C-weeks) accumulated over a 12-week period15,16, or sometimes using monthly SST data to compute Degree Heating Months (DHM in °C-months) over 3 months17,18. Numerous studies have documented coral bleaching in shallow water linked to periods of anomalously high SST and accumulated DHW, especially during El Niño events1,2,4,19,20,21. However, DHW generally only explain a limited proportion of the observed variation in bleaching, even for communities in very shallow water (e.g., 50% of bleaching variation at 2 m depth7). A number of studies have documented bleaching that was less than that predicted based on contemporaneous SST and DHW. This discrepancy has sometimes been hypothesised to reflect ongoing coral acclimatisation to increasing temperatures and/or shifts in community composition towards more heat-tolerant genotypes and species22,23,24,25,26. Bleaching rates higher than predicted by SST, under limited DHW, have also been documented, and can be species-specific and more pronounced in heat-intolerant cryptic species27. It is unclear to what extent the relatively limited power of SST metrics such as DHW to predict coral bleaching results from an incomplete description of environmental conditions, a lack of nuance in describing biological thresholds and organisms’ reactions to elevated temperatures, genetic variation and cryptic species27,28, or, most likely, a combination of factors.Estimates of surface MHW severity are also highly sensitive to both the spatial and temporal scales of SST data considered. Present-day satellite products allow degree heating to be calculated at relatively fine spatial and temporal resolution using SST data measured daily over pixels of ~25 km2 size (e.g., NOAA Coral Reef Watch29). While many heating assessments in the context of coral bleaching continue to focus on long-term heating based on temperature anomalies accumulated across months (12 weeks or 3 months for DHW or DHM, respectively)18,20,30, here we focus on higher-resolution heating calculated as Degree Heating Days (DHD in °C-days) using daily SST data. Calculation of DHD over 12-day windows is analogous to the more coarse resolution DHW (i.e., weekly data over 12 weeks)31 but with different units, finer temporal resolution, and shorter time lags between the actual elevation of environmental temperatures and the resulting accumulated heating metric (see the DHD and DHW comparison in Fig. S5).To investigate the role of spatial scale in characterising MHWs, we analysed SST at a range of scales around Moorea between 1985 and 2019: 2° × 2° (~50,000 km2), 1° × 1° (~12,300 km2) and 0.1° × 0.1° (~100 km2) (see Fig. S1). The importance of spatial scales in heating severity apparent at the surface is reinforced by the regional heterogeneity in SST (see Fig. 2 and data animations in the Supplementary Information). Considering heating at local scales and finer temporal resolutions maximises the potential to detect, characterise and compare heating events at the surface in a way that is relevant to in situ MHW conditions (see also Guo et al.32. for the importance of scales in MHW assessments). For instance, when assessed using NOAA’s ‘Regional Bleaching Heat Stress Gauges’ for the Society Islands, DHW reached 4.54 °C-weeks in April 2016 and 5.35 °C-weeks in May 201933, indicating moderate likelihood of bleaching during both events, although inherent to calculations of accumulated DHW, the maximum heating was centred multiple weeks after the in situ heating events actually occurred (e.g., see Fig. S5e, f). Heat accumulation using higher resolution DHD across 12-day windows around Moorea itself (2° × 2°) was more closely aligned temporally to the heating events, suggesting a much higher bleaching risk in 2016 (6.83 °C-days) than 2019 (2.10 °C-days; Table S3). This likely reflects the localised heterogeneity in SST during the 2019 MHW, when hotter surface conditions prevailed north of Moorea (Fig. 1b). At a local scale, assessed using SST within a ~10 × 10 km area north of Moorea (see Fig. S1), heating was similar to regional estimates in 2016 (5.73 °C-days), but in 2019 revealed an intense, localised heatwave over Moorea’s north shore (15.4 °C-days; see Fig. 1e, f; Table S1). Because of these demonstrated advantages of higher temporal-resolution analysis, we rely on DHD rather than DHW for our analysis of MHW patterns among years at Moorea.Fig. 2: Regional sea-surface temperature (SST) variability during the peak of the 2016 and 2019 surface marine heatwaves around Moorea suggested hotter conditions during 2016.Panels show SST over a, b 20° × 20° and c, d 10° × 10° during 2016 and 2019, respectively, focusing on the date of peak SST observed over Moorea’s north shore (8 April 2016 and 4 April 2019). Dashed squares in (a, b) show extent of (c, d) and those in (c, d) the extent of data shown in Fig. 1a, b (2° × 2°). Coastlines based on Wessel and Smith77.Full size imageContrasting surface and subsurface heatingComparing surface and subsurface MHWs is challenging for many coastal ecosystems, even once SST data of appropriately fine spatial and temporal scale are obtained, due to a lack of long-term, in-situ temperature data through which to assess mean climatological patterns below the sea surface. Moorea represents one of the few coral reef systems with consistent in-situ observations over timescales (decades) and depths (sea surface to 40 m) relevant to understanding the oceanographic drivers of subsurface heating and their impacts on coral bleaching. Our analysis of long-term SST records indicates there have been 16 local-scale (0.1° × 0.1°) surface MHWs over the north shore of Moorea relative to a MMM-based bleaching threshold of 29.8 °C (Table S4), compared to 14 regional-scale (2° × 2°) events (Table S3). Localised heating over the north shore was often greater than regional SST would suggest, with the hottest event recorded in 2003 reaching 17.7 °C-days locally (Table S4), compared to 15.5 °C-days regionally (Table S3). Further details on historical events can be found in ‘Surface MHW history around Moorea’ in the Supplementary Information, which provides context for the six more recent surface MHWs that have occurred over the north shore since continuous in-situ, reef-level observations began at Moorea in 2005 (MHWs in 2007, 2012, 2015, 2016, 2017, and 2019; Table S4). Two recent, contrasting events in 2016 and 2019 demonstrate the extent to which thermal environments at depth, and the associated severity of coral bleaching, can vary substantially from predictions based on sea-surface conditions.MHW severity based only on SST can miss important information on the conditions experienced by organisms at depths greater than the surface skin layer quantified through remote sensing, which may only be few millimetres thick34. For example, although the localised peak in sea-surface temperatures were essentially identical between 2016 (30.1 °C; Fig. 1a) and 2019 (30.2 °C; Fig. 1b), and regional surface heating metrics and warnings were similar33, markedly different heat accumulation occurred due to the different duration that temperatures remained above the putative coral bleaching threshold (MMM + 1 = 29.8 °C; up to 2 days in April 2016 compared to up to 11 days in April 2019; Table S1). Yet these results from local SST—of similar SST maximums in 2016 and 2019, but longer durations above the threshold in 2019—only capture some of the significant differences that led to constating MHW severity and ecological outcomes between years and across depths.Daily average temperatures measured in situ at reef level in water depths of 10–40 m over ~15 years are well correlated with daily SST (r2 = 0.94–0.78 at 10–40 m). However, the strength of the relationship between SSTs and in situ temperatures declines with increasing water depth, even when in situ temperatures are averaged to a daily resolution31,35. The potential for subsurface attenuation of heating over coral reefs has previously been demonstrated using high-resolution in-situ water temperature data in the context of both regional upwelling and local internal-wave climates31,36,37, with observations of periodic transport of deeper, cooler water onto reef habitats at a large number of reefs globally31,35,38. The propagation of internal-wave energy is associated with significant vertical displacements of density isopycnals and isotherms; e.g., ~60 m displacements along the Hawaiian Ridge39. Upon encountering a sloping bottom, internal-wave dynamics become complex, and, for habitats on the fore reef slope, typically result in rapid, periodic cooling (rather than oscillations around a mean temperature) as water masses associated with e.g., 24–27 °C isotherms are vertically advected onto the reef and recede again31,40,41. In deeper reef habitats there may also be periodic heating associated with exposure to warmer surface water masses when internal waves lead to downward displacement of isotherms31, but the overall magnitude of any resulting net heating is small across the depth range considered here (i.e., no average heating at depths of 40 m and less; Fig. S4).To separate the effects of low- and high-frequency processes driving heating across the reef slope, we used a filtering approach specifically designed and validated by Wyatt et al.31. to estimate coral reef thermal regimes without internal waves. In situ temperatures were filtered to isolate variability at frequencies higher and lower than the local inertial period (~40.0 h), effectively removing the effects of internal waves from lower frequency processes (i.e., multi-day weather patterns and seasonal effects; see Fig. S3). Contrasting the observed and filtered in-situ temperature variations (black and white lines, respectively, in Fig. 3) highlights differences in the processes driving the 2016 and 2019 subsurface MHWs. During the 2019 MHW around Moorea, the filtered, or ‘non-internal wave’ (NIW), temperatures closely resembled the observed temperatures (Fig. 3e–h, m–p) implying limited internal-wave cooling (IWC). Consistent warming across the water column was evident in 2019 and temperatures remained above the coral bleaching threshold for multiple days during early to late April (Table S1). By contrast, the 2016 MHW was characterized by temperatures remaining generally below the bleaching threshold and significant high-frequency variability indicative of IWC across depths, such that temperatures only exceed the predicted bleaching threshold for hours or less at a time (Fig. 3a–d, i–l; Table S1). The high-resolution temperature observations show that IWC was greatly reduced during 2019 (Fig. 3e–h). The power spectral density of observed temperatures, concentrated at semi-diurnal frequencies and consistent with internal-wave forcing at this location41, was significantly higher in 2016 and lower in 2019 than the average across years at 10 m depth (Fig. 4a; see inset). In deeper water at 20–40 m depths, temperature variance within the semi-diurnal frequency band increased relative to shallow depths and became more similar between the two events, such that at 40 m the semi-diurnal variability was equivalent in 2016 and 2019 (Fig. 4b–d; see insets). However, this similarity in temperature variance does not indicate an equivalent magnitude of IWC, since variability during 2019 (Fig. 3p) was around a warmer background temperature closer to the coral bleaching threshold. Extending the comparison of 2016 and 2019 to other recent local MHWs demonstrates two distinct types of events: greater IWC across reef depths during 2012, 2015, and 2016 MHWs, versus reduced IWC during the 2007, 2017 and 2019 MHWs (Fig. 5).Fig. 3: Contrasting reef-level temperature variations across depths on Moorea’s north shore during the 2016 and 2019 marine heatwaves.Panels on the left show the observed high-frequency water temperature variations (black lines, measured at 2-min intervals) during the hottest months (Apr–May) in a–d 2016 and e–h 2019 at a, e 10, b, f 20, c, g 30 and d, h 40 m depths. Right panels focus on relative variation during the heatwave peaks across the same depths: i–l* 06–12 Apr 2016 and m–p 11–17 Apr 2019. Non-internal-wave temperature variations are shown based on observed temperatures filtered to remove the high-frequency influence of internal waves (white lines). The satellite-derived sea-surface temperatures (SST; grey line) are shown for comparison to in situ temperatures. The horizontal dashed line shows the ‘bleaching threshold’ (maximum monthly mean + 1 °C) and the background shading provides a reference relative to temperatures above (red), equal (yellow) and below (blue) this threshold. The red dashed squares denote the axis limits in the right panels. *Note: 40 m logger during 2016 incorrectly recorded at 2-h interval.Full size imageFig. 4: Reduced semi-diurnal temperature variability during the summer of 2019 in shallower water on the north shore of Moorea.Power spectral density (PSD) plots (logarithmic scale) were computed within a 12-day window at a 10 m, b 20 m, c 30 m, and d 40 m water depths during the summer months (Dec–May) in 2016 (blue), 2019 (red), and 2004–2018 (black; excluding 2016 and 2019). Shading shows the 95% confidence intervals for each PSD. The tidal constituents (dotted lines) show variance consistent with semi-diurnal (M2) forcing across depths, with diurnal (K1) forcing in 10 m of water along with some variability consistent with the shallow water lunar overtide (M4). Insets show details of semi-diurnal differences at each depth.Full size imageFig. 5: Comparison of internal-wave cooling (IWC) across depths during recent surface marine heatwaves (MHWs) around Moorea.Based on average a daily temperature variance (in °C2) and b IWC ((overline{{{{{{rm{IWC}}}}}}}) in °C) during the six recent local MHWs (see Table S4 for dates) that can be grouped into: (1) high IWC events during 2012 (blue), 2015 (green) and 2016 (purple); and, (2) low IWC events that coincided with bleaching events in 2007 (red) and 2019 (orange), along with early 2017 (yellow). The daily variance and (overline{{{{{{rm{IWC}}}}}}}) during Dec–Apr across all years (2005–2019) is shown for reference (black dashed lines). Contours of c–f heat accumulation as degree heating days (DHD in °C days) and g–j degree cooling days due to internal waves (DCDIW in °C days) across depths are shown for 2007, 2012, 2016 and 2019. Due to data gaps in the in situ records, contours are not shown for the 2015 or 2017 events.Full size imageThe magnitude of temperature fluctuations produced by IWC, i.e., occurring at the semi-diurnal frequency, are of similar magnitude to long-term ocean warming and climate change threatening coral reefs globally. The average IWC ((overline{{{{{{rm{IWC}}}}}}})) during the high-IWC MHWs (2012, 2015 and 2016) was between 0.14 and 0.60 °C (Fig. 3a; Table S4) and comparable to the overall SST increase measured over tropical coral reefs during the last four decades (~0.65 °C18). As a result of the subsurface cooling caused by internal waves, the 2016 MHW, which was moderate at the surface (5.7 °C-days), was mild at 10 m ( More

Field approachOur searches uncovered 301 papers reporting field studies. After screening the titles abstracts, and full texts, we kept 130 articles for the analysis (Supplementary Fig. 1), from which we obtained 1342 observations regarding 57 Cx. mosquito species from 28 countries and 135 localities (Fig. 1A). Of these 1342 observations, 733 (54.61%) were classified as high quality, (i.e., the number of individuals tested was specified) (Supplementary Tables 1 and 2). The best represented countries were the USA (64.7%, number of observations = 869), Italy (9.3%, n = 125), and Iran (2.9%, n = 39). Based on mosquito field surveillance and individuals testing positive, we concluded that JES is distributed mainly in the Nearctic, Palearctic and Oriental regions (Fig. 1A).Figure 1(A) Weighted Minimum Infection Rates and (B) Weighted Transmission Efficiency of mosquito populations for JES. High-quality data. The size of circles represents the magnitude of the estimates. Map was generated using R software version 4.1.2 with the packages mapdata, maps and tydiverse (https://www.r-project.org) and edited with Inkscape (https://inkscape.org/es/).Full size imageWest Nile virusWNV was detected mainly in the USA (76.5%, number of observations = 826), Italy (4.9%, n = 53) and Iran (3.6%, n = 39) (Fig. 1A). We also recorded 23 species (57%, 41 species) interacting with this virus (Supplementary Tables 1 and 3). Cx. quinquefasciatus became naturally infected in North America [Infection Frequency (IF) = 2.33] (Table 1). We recorded WNV interacting with Cx. tritaeniorhynchus in Asia (IF = 1.02), with Cx. pipiens in Europe (IF = 1.74) and with Cx. antennatus, Cx. neavei, Cx. perexiguus, Cx. perfuscus, Cx. poicilipes, Cx. quinquefasciatus and Cx. tritaeniorhynchus in Africa (IF = 1) (Supplementary Table 3).Table 1 Description of the variables.Full size tableThe highest infection rates were found in North America in Cx. restuans [Standardized minimum infection rate (SMIR) = 56.01], and in Africa and Europe in Cx. pipiens (SMIR = 20.45 and 29.25, respectively). No positive SMIR values were reported in Asian mosquitoes, and Oceanic mosquitoes were not sampled for this virus (Fig. 2A and Supplementary Table 3). The highest infection risk or potential was recorded in species from the USA, such as Cx. restuans (Infection Risk (IR) = 69.50), Cx. pipiens (IR = 55) and Cx. tarsalis (IR = 52.16) (Fig. 3A and Supplementary Table 3). Finally, WNV lineage 1 was detected in Algeria, Turkey, Portugal, Mexico, Tunisia, Iran, Spain and Italy, lineage 2 in Italy, Bulgaria, Greece and the Czech Republic, and lineage 5 in India (Supplementary Table 2).Figure 2(A) Box plots for the Weighted Minimum Infection Rates and (B) Weighted Transmission Efficiency Rates for JES. Boxes indicate 2nd and 3rd quartiles, vertical lines upper and lower quartiles, and horizontal lines the median. Points indicate outliers. The Y axis was transformed to Sqrt (Square root).Full size imageFigure 3JES (A) Infection Risk and (B) Transmission Risk by mosquito species.Full size imageJapanese encephalitis virusJEV was detected mainly in Taiwan (21.6%, n = 24), Korea (18%, n = 20) and Australia (15.3%, n = 17) (Fig. 1A). We found 23 mosquito species interacting with JEV: Cx. vishnui was the one most frequently found to be positive (IF = 1.20), followed by Cx. tritaeniorhynchus (IF = 1.17), Cx. pipiens and Cx. annulus (IF = 0.98) in Asia, while the most susceptible species in Oceania were Cx. sitiens and Cx. gelidus (IF = 0.71) (Supplementary Table 3).The highest SMIR values were recorded in Asia in Cx. rubithoracis (62.38), Cx. annulus (47.68) and Cx. tritaeniorhynchus (28.16) (Fig. 2A and Supplementary Table 3), and Cx. fuscocephala had the highest estimated natural IR (Fig. 3A, Supplementary Table 3). Three genotypes were recorded: genotype I (strain VNKT/479/2007, VNKT/486/2007, and JEV Ishikawa12), genotype III (Tibet-Culex-JEV1-5), and genotype V (K12YJ1174). These were isolated in China, Vietnam, and Japan (Genotype I), Italy, China (Genotype III), and Korea (Genotype V) (Supplementary Table 2).Usutu virusField studies on USUV have been conducted in Europe, Africa, and Asia, most of them in Italy (66.6%, n = 72), Czechoslovakia (11.1%, n = 12) and Slovakia (7.4%, n = 8) (Fig. 1A). Six species were reported to be susceptible to natural infection. Cx. perexiguus had the highest IF and SMIR (1.30) (Supplementary Table 3). In Africa, Cx. antennatus (IF = 1), and in Asia Cx. pipiens (IF = 1) were the most likely to be positive, while Cx. pipiens had the highest IR (5.19) (Fig. 3A, Supplementary Table 3). The recorded strains were USU181_09/USU090-10/USU173_09 (Italy) and USU/Croatia/Zagreb-102/2018 (Italy).St. Louis encephalitis virusThe field studies on SLEV focused on North America (97.7%, n = 43) and Brazil (2.2%, n = 1). Three species were recorded interacting with this virus. Cx. erraticus had the highest IF (2.06), SMIR (2.06) and IR, followed by Cx. quinquefasciatus (North America) (IF = 0.73, SMIR = 1.97) (Fig. 2A, Supplementary Table 3).The highest estimated IR of JES was for Cx. pipiens (Europe), which can be naturally infected with WNV and USUV, followed by Cx. quinquefasciatus (North America), which can be infected with WNV and SLEV (Fig. 3A).Experimental approachExperimental studies were reported in 481 articles. After screening the titles, abstracts, and full texts, as well as opportunistic records, 95 articles remained for the analysis (Supplementary Fig. 2). From these we obtained 189 high quality observations of the TE of JES in 11 countries, 40 localities, and 21 species (Fig. 1B, Supplementary Table 1). The USA was the best represented country (54.4%, n = 103), followed by Germany (13.2%, n = 25) and Australia (12.6%, n = 24). There was, however, a notable lack of information on the vector competence of Cx. mosquitos for JES in many regions of the world, such as Central and South America, and Africa (Fig. 1B).The most common means of infection was oral (94.8%, 395 observations), while the rest were intrathoracic. Intrathoracic infection bypasses the midgut barrier so is not considered natural infection. We therefore carried out the subsequent analyses using only the data on oral infection (Supplementary Table 4).We used a generalised linear model (GLM) for the statistical analysis, which was conducted only on the WNV dataset (strain NY99), the only one with sufficient observations for the purpose (n = 63). We did not find a significant effect of viral titre, temperature, or days post infection on TE. However, more data with a wide range of values is necessary to confirm these observations. On the other hand, we found that the Extrinsic Incubation Period (as DPI) was shorter at higher temperatures (Fig. 4 and Supplementary Table 5).Figure 4Relationship between temperature and Days Post Infection for WNV strain NY99.Full size imageWest Nile virusMosquito populations from many locations on all continents have been studied for their vector competence for this virus, particularly in the USA (60.3% of observations, n = 96), Germany (15.7%, n = 25) and Australia (6.9%) (Fig. 1B). Our bibliographic research revealed 21 species of Cx. with the ability to transmit WNV under laboratory conditions (Supplementary Table 6). Cx. pipiens (North America) and Cx. tarsalis were the most frequently studied species and were the most efficient in transmitting the virus (Transmission Frequency (TF) = 2.33) (Table1). Cx. quinquefasciatus had the highest TF (1.70) in Africa, Cx. modestus in Europe (TF = 1.32), and Cx. annulirostris and Cx. quinquefasciatus in Oceania (TF = 1.48) (Supplementary Table 6).Concerning Standardized Transmission Rates (STE) estimates, Cx. quinquefasciatus had the highest values in the USA (STE = 1.63), Cx. pipiens in Europe (0.90), Cx. tritaeniorhynchus in Asia (1.8), Cx. neavei in Africa (0.17) and Cx. annulirostris in Oceania (2.45) (Fig. 2B, Supplementary Table 6). We found 20 different strains of WNV tested. The TE of the various WNV strains vary considerably, but lineage 1 was more efficient than lineage 2. There were also more studies on the lineage 1 strains (n = 11), which exhibited high variation (Fig. 5).Figure 5Box plots for WNV (A) lineages and (B) strains used to measure Weighted Transmission Rates.Full size imageJapanese encephalitis virusJEV has been studied mainly in mosquito populations from France (45%, n = 20) and Australia (34%, n = 15), but also the United Kingdom, India, Taiwan, New Zealand, and the USA (Fig. 1B). Six mosquito species are capable of transmitting JEV. Cx. pipiens (Europe) had the highest TF (1.85), while Cx. gelidus had high values of STE (1.73) (Fig. 3B and Supplementary Table 6).St. Louis encephalitis virusVector competence for SLEV has been studied in two countries: the USA (93.3%, n = 42) and Argentina (6.6%, n = 3), and 7 mosquito species have been investigated. Cx. nigripalpus was the most efficient in transmitting the virus (TF = 1.60), while Cx. pipiens had the highest STE (0.68) (Figs. 2B, 3B and Supplementary Table 6).Usutu virusStudies have also been conducted on the Usutu virus in mosquito populations in the USA (28.57%, n = 4), the United Kingdom (42.8%, n = 6) and Senegal (25%, n = 4), in particular on Cx. neavei, Cx. pipiens and Cx. quinquefasciatus (TF = 1). Cx. neavei also had the highest STE (0.79) (Fig. 3B).We found reports of JES transmission under laboratory conditions in 22 Cx. species, and natural infections in 32 species (55.1% of the total sample) in the field. Cx. pipiens complex (biotypes quinquefasciatus, pipiens, molestus and pallens) was the most common vector accounting for 36.9% (n = 660) of the experimental observations and 25.7% (n = 1342) of the field observations. With both approaches, WNV was the most common flavivirus, accounting for 80.4% of the field observations and 86.7% of the experimental data (Fig. 1A,B). Only WNV, therefore, had enough observations to make comparison between the experimental and field data possible. We were able to compare 16 mosquito species and found a high positive correlation between TF and IF (R = 0.57, p = 0.02) (Fig. 7).In summary, we found that the species with the highest infection-transmission risk (IRT) for WNV was Cx. restuans, for USUV it was Cx. pipiens (Europe), for SLEV Cx. quinquefasciatus (North America), and for JEV Cx. gelidus (Oceania) (Fig. 6 and Supplementary Tables 2 and 6).Figure 6JES infection-transmission risk by continent and flavivirus.Full size image More

Feit, B. et al. Landscape complexity promotes resilience of biological pest control to climate change. Proc. R. Soc. B. 288, 20210547 (2021).Article 

Google Scholar 
Hall-Spencer, J. M. & Harvey, B. P. Ocean acidification impacts on coastal ecosystem services due to habitat degradation. Emerg. Top. Life Sci. 3, 197–206 (2019).Article 
CAS 

Google Scholar 
Loke, L. H. L. & Todd, P. A. Structural complexity and component type increase intertidal biodiversity independently of area. Ecology 97, 383–393 (2016).Article 

Google Scholar 
Oliver, T. H. et al. Biodiversity and resilience of ecosystem functions. Trends in Ecol. Evol. 30, 673–684 (2015).Article 

Google Scholar 
Bullock, J. M. et al. Future restoration should enhance ecological complexity and emergent properties at multiple scales. Ecography ecog. 4, 05780 (2022).Ortega, J. C. G., Thomaz, S. M. & Bini, L. M. Experiments reveal that environmental heterogeneity increases species richness, but they are rarely designed to detect the underlying mechanisms. Oecologia 188, 11–22 (2018).Article 

Google Scholar 
Griffin, J. N., Byrnes, J. E. K. & Cardinale, B. J. Effects of predator richness on prey suppression: a meta-analysis. Ecology 94, 2180–2187 (2013).Article 

Google Scholar 
Katano, I., Doi, H., Eriksson, B. K. & Hillebrand, H. A cross-system meta-analysis reveals coupled predation effects on prey biomass and diversity. Oikos 124, 1427–1435 (2015).Article 

Google Scholar 
Loke, L. H. L., Ladle, R. J., Bouma, T. J. & Todd, P. A. Creating complex habitats for restoration and reconciliation. Ecol. Eng. 77, 307–313 (2015).Article 

Google Scholar 
Torres-Pulliza, D. et al. A geometric basis for surface habitat complexity and biodiversity. Nat. Ecol. Evol. 4, 1495–1501 (2020).Article 

Google Scholar 
Chesson, P. Mechanisms of maintenance of species diversity. Annu. Rev. Ecol. Syst. 31, 343–366 (2000).Article 

Google Scholar 
Chesson, P. & Kuang, J. J. The interaction between predation and competition. Nature 456, 235–238 (2008).Article 
CAS 

Google Scholar 
Terborgh, J. W. Toward a trophic theory of species diversity. Proc. Natl. Acad. Sci. USA 112, 11415–11422 (2015).Article 
CAS 

Google Scholar 
Pringle, R. M. et al. Predator-induced collapse of niche structure and species coexistence. Nature 570, 58–64 (2019).Article 
CAS 

Google Scholar 
Sandom, C. et al. Mammal predator and prey species richness are strongly linked at macroscales. Ecology 94, 1112–1122 (2013).Article 

Google Scholar 
Grabowski, J. H. Habitat complexity disrupts predator-prey interactions but not the trophic cascade on oyster reefs. Ecology 85, 995–1004 (2004).Article 

Google Scholar 
Crowder, L. B. & Cooper, W. E. Habitat structural complexity and the interaction between bluegills and their prey. Ecology 63, 1802 (1982).Article 

Google Scholar 
Almany, G. R. Does increased habitat complexity reduce predation and competition in coral reef fish assemblages? Oikos 106, 275–284 (2004).Article 

Google Scholar 
Anderson, T. L. & Semlitsch, R. D. Top predators and habitat complexity alter an intraguild predation module in pond communities. J. Anim. Ecol. 85, 548–558 (2016).Article 

Google Scholar 
Brothers, C. A. & Blakeslee, A. M. H. Alien vs predator play hide and seek: How habitat complexity alters parasite mediated host survival. J. Exp. Mar. Biol. Ecol. 535, 151488 (2021).Article 

Google Scholar 
Horinouchi, M. et al. Seagrass habitat complexity does not always decrease foraging efficiencies of piscivorous fishes. Mar. Ecol. Prog. Ser. 377, 43–49 (2009).Article 

Google Scholar 
Ryer, C., Stoner, A. & Titgen, R. Behavioral mechanisms underlying the refuge value of benthic habitat structure for two flatfishes with differing anti-predator strategies. Mar. Ecol. Prog. Ser. 268, 231–243 (2004).Article 

Google Scholar 
Flynn, A. J. & Ritz, D. A. Effect of habitat complexity and predatory style on the capture success of fish feeding on aggregated prey. J. Mar. Biol. Ass. 79, 487–494 (1999).Article 

Google Scholar 
Klecka, J. & Boukal, D. S. The effect of habitat structure on prey mortality depends on predator and prey microhabitat use. Oecologia 176, 183–191 (2014).Article 

Google Scholar 
James, P. L. & Heck, K. L. The effects of habitat complexity and light intensity on ambush predation within a simulated seagrass habitat. J. Exp. Mar. Biol. Ecol. 176, 187–200 (1994).Article 

Google Scholar 
Michel, M. J. & Adams, M. M. Differential effects of structural complexity on predator foraging behavior. Behav. Ecol. 20, 313–317 (2009).Article 

Google Scholar 
Preisser, E. L., Bolnick, D. I. & Benard, M. F. Scared to death? The effects of intimidation and consumption in predator-prey interactions. Ecology 86, 501–509 (2005).Article 

Google Scholar 
Preisser, E. L., Orrock, J. L. & Schmitz, O. J. Predator hunting mode and habitat domain alter nonconsumptive effects in predator-prey interactions. Ecology 88, 2744–2751 (2007).Article 

Google Scholar 
Rypstra, A. L., Schmidt, J. M., Reif, B. D., DeVito, J. & Persons, M. H. Tradeoffs involved in site selection and foraging in a wolf spider: effects of substrate structure and predation risk. Oikos 116, 853–863 (2007).Article 

Google Scholar 
Janssen, A., Sabelis, M. W., Magalhães, S., Montserrat, M. & van der Hammen, T. Habitat structure affects intraguild predation. Ecology 88, 2713–2719 (2007).Article 

Google Scholar 
Grabowski, J. H., Hughes, A. R. & Kimbro, D. L. Habitat complexity influences cascading effects of multiple predators. Ecology 89, 3413–3422 (2008).Article 

Google Scholar 
Hughes, A. R. & Grabowski, J. H. Habitat context influences predator interference interactions and the strength of resource partitioning. Oecologia 149, 256–264 (2006).Article 

Google Scholar 
Bonett, D. G. Meta-analytic interval estimation for standardized and unstandardized mean differences. Psychol. Methods 14, 225–238 (2009).Article 

Google Scholar 
Huey, R. B. & Pianka, E. R. Ecological consequences of foraging mode. Ecology 62, 991–999 (1981).Article 

Google Scholar 
Egger, M., Smith, G. D., Schneider, M. & Minder, C. Bias in meta-analysis detected by a simple, graphical test. BMJ 315, 629–634 (1997).Article 
CAS 

Google Scholar 
Ritchie, E. G. & Johnson, C. N. Predator interactions, mesopredator release and biodiversity conservation. Ecol. Lett. 12, 982–998 (2009).Article 

Google Scholar 
Chaplin-Kramer, R., O’Rourke, M. E., Blitzer, E. J. & Kremen, C. A meta-analysis of crop pest and natural enemy response to landscape complexity: pest and natural enemy response to landscape complexity. Ecol. Lett. 14, 922–932 (2011).Article 

Google Scholar 
Paxton, A. B. et al. Meta-analysis reveals artificial reefs can be effective tools for fish community enhancement but are not one-size-fits-all. Front. Mar. Sci. 7, 282 (2020).Article 

Google Scholar 
Eggleston, D. B., Lipcius, R. N., Miller, D. L. & Coba-Cetina, L. Shelter scaling regulates survival of juvenile Caribbean spiny lobster Panulirus argus. Mar. Ecol. Prog. Ser. 62, 79–88 (1990).Rogers, A., Blanchard, J. L. & Mumby, P. J. Fisheries productivity under progressive coral reef degradation. J. Appl. Ecol. 55, 1041–1049 (2018).Article 

Google Scholar 
Gontijo, L. M. Engineering natural enemy shelters to enhance conservation biological control in field crops. Biol. Control 130, 155–163 (2019).Article 

Google Scholar  More

Martínez, G. M. & Renno, N. O. Water and brines on Mars: Current evidence and implications for MSL. Sp. Sci. Rev. 175(1), 29–51 (2013).Article 
ADS 

Google Scholar 
Orosei, et al. Radar evidence of subglacial liquid water on Mars. Science 361(6401), 490–493. https://doi.org/10.1126/science.aar7268 (2018).Article 
ADS 

Google Scholar 
Mikucki, J. A. et al. Deep groundwater and potential subsurface habitats beneath an Antarctic dry valley. Nat. Commun. 6(6831), 1–9 (2015).
Google Scholar 
Forte, E., Dalle Fratte, M., Azzaro, M. & Guglielmin, M. Pressurized brines in continental Antarctica as a possible analogue of Mars. Sci. Rep. 6, 33158 (2016).Article 
ADS 

Google Scholar 
Siegert, M. J., Kennicutt, M. C. & Bindschadler, R. A. Antarctic Subglacial Aquatic Environments (Wiley, 2013).
Google Scholar 
Boulton, G. S., Caban, P. E. & van Gijssel, K. Groundwater flow beneath ice sheets: Part I—Large-scale patterns. Quatern. Sci. Rev. 14, 545–562 (1995).Article 
ADS 

Google Scholar 
Fricker, H. A., Carter, S. P., Bell, R. E. & Scambos, T. Active lakes of Recovery Ice Stream, East Antarctica: A bedrock-controlled subglacial hydrological system. J. Glaciol. 60(223), 1015–1030. https://doi.org/10.3189/2014JoG14J063 (2014).Article 
ADS 

Google Scholar 
Siegert, M. J. A wide variety of unique environments beneath the Antarctic ice sheet. Geology 44(5), 399–400. https://doi.org/10.1130/focus052016.1 (2016).Article 
ADS 
MathSciNet 

Google Scholar 
Lyons, W. B. et al. The geochemistry of englacial brine from Taylor Glacier, Antarctica. J. Geophys. Res. Biogeosci. 124, 633–648. https://doi.org/10.1029/2018JG004411 (2019).Article 

Google Scholar 
Campbell, S., Courville, Z., Sinclair, S. & Wilner, J. Brine, englacial structure and basal properties near the terminus of McMurdo Ice Shelf, Antarctica. Ann. Glaciol. 58, 74. https://doi.org/10.1017/aog.2017.26 (2017).Article 

Google Scholar 
Greene, S. et al. Canadian Shield brine from the Con Mine, Yellowknife, NT, Canada: Noble gas evidence for an evaporated Palaeozoic seawater origin mixed with glacial meltwater and Holocene recharge. Geochim. Cosmochim. Acta 72, 4008–4019. https://doi.org/10.1016/j.gca.2008.05.058 (2008).Article 
ADS 

Google Scholar 
Siegfried, M. R., Fricker, H. A., Carter, S. P. & Tulaczyk, S. Episodic ice velocity fluctuations triggered by a subglacial flood in West Antarctica. Geophys. Res. Lett. 43, 2640–2648. https://doi.org/10.1002/2016GL067758 (2016).Article 
ADS 

Google Scholar 
Stearns, L. A., Smith, B. E. & Hamilton, G. S. Increased flow speed on a large East Antarctic outlet glacier caused by subglacial floods. Nat. Geosci. 1(12), 827–831. https://doi.org/10.1038/ngeo356 (2008).Article 
ADS 

Google Scholar 
Kennicutt, M. C. et al. A roadmap for Antarctic and Southern Ocean science for the next two decades and beyond. Antarct. Sci. 27(01), 3–18. https://doi.org/10.1017/S0954102014000674 (2015).Article 
ADS 

Google Scholar 
Welch, K. A. et al. Spatial variations in the geochemistry of glacial meltwater streams in the Taylor Valley, Antarctica. Antarct. Sci. 22(06), 662–672. https://doi.org/10.1017/S0954102010000702 (2010).Article 
ADS 

Google Scholar 
Skidmore, M., Tranter, M., Tulaczyk, S. & Lanoil, B. Hydrochemistry of ice stream beds—evaporitic or microbial effects?. Hydrol. Process. 24(4), 517–523 (2010).
Google Scholar 
Lüttge, A. & Conrad, P. G. Direct observation of microbial inhibition of calcite dissolution. Appl. Environ. Microbiol. 20, 1627–1632 (2004).Article 
ADS 

Google Scholar 
Mikucki, J. A. & Priscu, J. C. Bacterial diversity associated with Blood Falls, a subglacial outflow from the Taylor Glacier, Antarctica. Appl. Environ. Microbiol. 73(12), 4029–4039 (2007).Article 
ADS 

Google Scholar 
Mikucki, J. A. et al. A contemporary microbially maintained subglacial ferrous “Ocean”. Science 324(5925), 397–400. https://doi.org/10.1126/science.1167350 (2009).Article 
ADS 

Google Scholar 
Chua, M. J. et al. Genomic and physiological characterization and description of Marinobacter gelidimuriae sp. Nov., a psychrophilic, moderate halophile from Blood Falls, an Antarctic subglacial brine. FEMS Microbiol. Ecol. 94, fiy021 (2018).Article 

Google Scholar 
Murray, A. E. et al. Microbial life at −13 °C in the brine of an ice-sealed Antarctic lake. PNAS 109, 20626–20631. https://doi.org/10.1073/pnas.1208607109 (2012).Article 
ADS 

Google Scholar 
Borruso, L. et al. A thin ice layer segregates two distinct fungal communities in Antarctic brines from Tarn Flat (Northern Victoria Land). Sci. Rep. 8, 1–9 (2018).Article 

Google Scholar 
Papale, M. et al. Microbial assemblages in pressurized Antarctic brine pockets (Tarn Flat, Northern Victoria Land): A hotspot of biodiversity and activity. Microorganisms 7, 333 (2019).Article 

Google Scholar 
Azzaro, M. et al. The prokaryotic community in an extreme Antarctic environment: The brines of Boulder Clay lakes (Northern Victoria Land). Hydrobiologia 848, 1837–1857. https://doi.org/10.1007/s10750-021-04557-2 (2021).Article 

Google Scholar 
Lo Giudice, A. et al. Prokaryotic diversity and metabolically active communities in brines from two perennially ice-covered Antarctic lakes. Astrobiology 21, 551–565 (2021).Article 
ADS 

Google Scholar 
Sannino, C. et al. Intra-and inter-cores fungal diversity suggests interconnection of different habitats in an Antarctic frozen lake (Boulder Clay, Northern Victoria Land). Environ. Microbiol. 22, 3463–3477 (2020).Article 

Google Scholar 
Bratina, B. J., Stevenson, B. S., Green, W. J. & Schmidt, T. M. Manganese reduction by microbes from oxic regions of the lake vanda (Antarctica) water column. Appl. Environ. Microbiol. 64, 3791–3797 (1998).Article 
ADS 

Google Scholar 
Tregoning, G. S. et al. A halophilic bacterium inhabiting the warm, CaCl2-rich brine of the perennially ice-covered Lake Vanda, McMurdo Dry Valleys, Antarctica. Appl. Environ. Microbiol. 81, 1988–1995 (2015).Article 
ADS 

Google Scholar 
Kwon, M. et al. Niche specialization of bacteria in permanently ice-covered lakes of the McMurdo Dry Valleys, Antarctica. Environ. Microbiol. 19, 2258–2271 (2017).Article 

Google Scholar 
Forte, E., Azzaro, M. & Guglielmin, M. Evidence of an unprecedented water erosion and supraglacial-fluvial sedimentation on an Antarctic glacier in the Holocene. Sci. Total Environ. 20, 20 (2022).
Google Scholar 
Doran, P. T. et al. Radiocarbon distribution and the effect of legacy in lakes of the McMurdo Dry Valleys, Antarctica. Limnol. Oceanogr. 59(3), 811–826. https://doi.org/10.4319/lo.2014.59.3.0811 (2014).Article 
ADS 

Google Scholar 
Saccò, M. et al. Salt to conserve: A review on the ecology and preservation of hypersaline ecosystems. Biol. Rev. 96, 2828–2850 (2021).Article 

Google Scholar 
Ramoneda, J. et al. Importance of environmental factors over habitat connectivity in shaping bacterial communities in microbial mats and bacterioplankton in an Antarctic freshwater system. FEMS Microbiol. Ecol. 97, fiab044 (2021).Article 

Google Scholar 
Saxton, M. A. et al. Sulfate reduction and methanogenesis in the hypersaline deep waters and sediments of a perennially ice-covered lake. Limnol. Oceanogr. 66, 1804–1818 (2021).Article 
ADS 

Google Scholar 
Frey, B. et al. Microbial diversity in European alpine permafrost and active layers. FEMS Microbiol. Ecol. 92, fiw018. https://doi.org/10.1093/femsec/fiw018 (2016).Article 

Google Scholar 
Hu, W. et al. Characterization of the prokaryotic diversity through a stratigraphic permafrost core profile from the Qinghai-Tibet Plateau. Extremophiles 20, 337–349 (2016).Article 

Google Scholar 
Alekseev, I., Zverev, A. & Abakumov, E. Microbial communities in permafrost soils of Larsemann Hills, Eastern Antarctica: Environmental controls and effect of human impact. Microorganisms 8(8), 1202 (2020).Article 

Google Scholar 
Tian, R. et al. Small and mighty: Adaptation of superphylum Patescibacteria to groundwater environment drives their genome simplicity. Microbiome 8, 51 (2020).Article 

Google Scholar 
Bowman, J. P., McCammon, S. A., Rea, S. M. & McMeekin, T. A. The microbial composition of three limnologically disparate hypersaline Antarctic lakes. FEMS Microbiol. Lett. 183, 81–88 (2000).Article 

Google Scholar 
Aislabie, J. & Bowman J. P. “Archaeal Diversity in Antarctic Ecosystems.” Polar Microbiology: The Ecology, Biodiversity and Bioremediation Potential of Microorganisms in Extremely Cold Environments 31–59 (CRC Press, 2010).
Google Scholar 
Zhang, C. J. et al. Spatial and seasonal variation of methanogenic community in a river-bay system in South China. Appl. Microbiol. Biotechnol. 104, 4593–4603. https://doi.org/10.1007/s00253-020-10613-z (2020).Article 

Google Scholar 
Bapteste, E., Brochier, C. & Boucher, Y. Higher-level classification of the archaea: Evolution of methanogenesis and methanogens. Archaea 1, 353–363 (2005).Article 

Google Scholar 
Bowman, J. P. et al. Psychroflexus torquis gen. nov., sp. nov., a psychrophilic species from Antarctic sea ice, and reclassification of Flavobacterium gondwanense (Dobson et al. 1993) as Psychroflexus gondwanense gen. nov., comb. nov.. Microbiology 144, 1601–1609 (1998).Article 

Google Scholar 
Donachie, S. P., Bowman, J. P. & Alam, M. Psychroflexus tropicus sp. Nov., an obligately halophilic Cytophaga-Flavobacterium-Bacteroides group bacterium from an Hawaiian hypersaline lake. Int. J. Syst. Evol. Microbiol. 54, 935–940 (2004).Article 

Google Scholar 
Zhong, Z. P. et al. Psychroflexus salis sp. Nov. and Psychroflexus planctonicus sp. Nov., isolated from a salt lake. Int. J. Syst. Evol. Microbiol. 66, 125–131 (2016).Article 

Google Scholar 
Chun, J., Kang, J. Y. & Jahng, K. Y. Psychroflexus salarius sp. Nov., isolated from Gomso salt pan. Int. J. Syst. Evol. Microbiol. 64, 3467–3472 (2014).Article 

Google Scholar 
Yoon, J. H., Kang, S. J., Jung, Y. T. & Oh, T. K. Psychroflexus salinarum sp. Nov., isolated from a marine solar saltern. Int. J. Syst. Evol. Microbiol. 59, 2404–2407 (2009).Article 

Google Scholar 
Buzzini, P., Turchetti, B. & Yurkov, A. Extremophilic yeasts: The toughest yeasts around?. Yeast 35, 487–497 (2018).Article 

Google Scholar 
Coleine, C., Stajich, J. E. & Selbmann, L. Fungi are key players in extreme ecosystems. Trends Ecol. Evol. S0169–5347(22), 00025–00028 (2022).
Google Scholar 
Gonçalves, V. N. et al. Taxonomy, phylogeny and ecology of cultivable fungi present in seawater gradients across the Northern Antarctica Peninsula. Extremophiles 21, 1005–1015 (2017).Article 

Google Scholar 
Ogaki, M. B. et al. Cultivable fungi present in deep-sea sediments of Antarctica: Taxonomy, diversity, and bioprospecting of bioactive compounds. Extremophiles 24, 227–238 (2020).Article 

Google Scholar 
Wedin, M., Döring, H. & Gilenstam, G. Saprotrophy and lichenization as options for the same fungal species on different substrata: Environmental plasticity and fungal lifestyles in the Stictis-Conotrema complex. New Phytol. 164, 459–465 (2004).Article 

Google Scholar 
Sterflinger, K. Black yeasts and meristematic fungi: Ecology, diversity and identification. In Biodiversity and Ecophysiology of Yeasts. The Yeast Handbook (eds Péter, G. & Rosa, C.) 501–514 (Springer, 2006).Chapter 

Google Scholar 
Canini, F. et al. Growth forms and functional guilds distribution of soil Fungi in coastal versus inland sites of Victoria Land, Antarctica. Biology (Basel) 10, 320 (2021).
Google Scholar 
Vaniman, D. T. et al. Magnesium sulfate salts and the history of water on Mars. Nature 431, 663–665 (2004).Article 
ADS 

Google Scholar 
Gendrin, A. et al. Sulfates in martian layered terrains: The OMEGA/Mars Express view. Science 307, 1587–1591 (2005).Article 
ADS 

Google Scholar 
Carr, M. H. & Head, J. W. I. I. I. Geologic history of Mars. Earth Planet Sci. Lett. 294, 185–203 (2010).Article 
ADS 

Google Scholar 
Ojha, L. et al. Spectral evidence for hydrated salts in recurring slope lineae on Mars. Nat. Geosci. 8, 829–832 (2015).Article 
ADS 

Google Scholar 
Cragin, J. H., Gow, A. J. & Kovacs, A. Chemical fractionation of brine in the McMurdo Ice Shelf, Antarctica. CRREL Rep. 20, 83–86 (1983).
Google Scholar 
Frank, T. D. & Gui, Z. Cryogenic origin for brine in the subsurface of southern McMurdo Sound, Antarctica. Geology 38(7), 587–590. https://doi.org/10.1130/G30849.1 (2010).Article 
ADS 

Google Scholar 
Gardner, C. B. & Lyons, W. B. Modeled geochemical composition of cryogenically produced subglacial Brines, Antarctica. Antarct. Sci. 31(3), 165–166 (2019).Article 
ADS 

Google Scholar 
Lyons, W. B. et al. Halogen geochemistry of the McMurdo Dry Valleys lakes, Antarctica: Clues to the origin of solutes and lake evolution. Geochim. Cosmochim. Acta 69, 305–323 (2005).Article 
ADS 

Google Scholar 
Armienti, P. & Baroni, C. Cenozoic climatic change in Antarctica recorded by volcanic activity and landscape evolution. Geology 27(7), 617–620 (1999).Article 
ADS 

Google Scholar 
Di Nicola, L. et al. Multiple cosmogenic nuclides document complex Pleistocene exposure history of glacial drifts in Terra Nova Bay (northern Victoria Land, Antarctica). Quatern. Res. 71(1), 83–92 (2009).Article 
ADS 
MathSciNet 

Google Scholar 
Levy, R. et al. Late Neogene climate and glacial history of the Southern Victoria Land coast from integrated drill core, seismic and outcrop data. Glob. Planet. Change 80–81, 61–84 (2012).Article 
ADS 

Google Scholar 
Prebble, J. G., Raine, J. I., Barrett, P. J. & Hannah, M. J. Vegetation and climate from two Oligocene glacioeustatic sedimentary cycles (31 and 24 Ma) cored by the Cape Roberts Project, Victoria Land Basin, Antarctica. Palaeogeogr. Palaeoclimatol. Palaeoecol. 231, 41–57 (2006).Article 

Google Scholar 
Tedersoo, L. et al. Shotgun metagenomes and multiple primer pair barcode combinations of amplicons reveal biases in metabarcoding analyses of fungi. Myco Keys 10, 1–43 (2015).Article 

Google Scholar 
Andrews, S. FastQC: A quality control tool for high throughput sequence data. http://www.bioinformatics.babraham.ac.uk/projects/fastqc. (2010).Bolyen, E. et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat. Biotechnol. 37, 852–857. https://doi.org/10.1038/s41587-019-0209-9 (2019).Article 

Google Scholar 
Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583. https://doi.org/10.1038/nmeth.3869 (2016).Article 

Google Scholar 
Nilsson, R. H. et al. The UNITE database for molecular identification of fungi: Handling dark taxa and parallel taxonomic classifications. Nucleic Acids Res 47, D259–D264. https://doi.org/10.1093/nar/gky1022 (2019).Article 

Google Scholar  More

Study populationThe study is based on data from mothers and children participating in a longitudinal Southwest Finland cohort, Steps to Healthy development of Children (the STEPS Study) (described in detail in Lagström et al.31). The STEPS study is an ongoing population-based and multidisciplinary study that investigates children’s physical, psychological and social development, starting from pregnancy and continuing until adolescence. All Finnish- and Swedish-speaking mothers delivering a child between 1 January, 2008 and 31 March, 2010 in the Hospital District of Southwest Finland formed the cohort population (in total, 9811 mothers and their 9936 children). Altogether, 1797 mothers with 1805 neonates volunteered as participants for the intensive follow-up group of the STEPS Study. Mothers were recruited by midwives either during the first trimester of pregnancy while visiting maternity health care clinics, or after delivery on the maternity wards of Turku University Hospital or Salo Regional Hospital, or by a letter mailed to the mothers. The participating mothers differ slightly from the whole cohort population in some background characteristics (being older, with first-born child and higher socioeconomic status)31. The ethics committee of the Hospital District of Southwest Finland has approved the STEPS Study (2/2007) and all methods were performed in accordance with relevant guidelines and regulations. Written informed consent was obtained from all the participants and, for children, from one parent for study participation. Subjects have been and are free to withdraw from the study at any time without any specific reason. The STEPS Study have the appropriate government authorization to the use of the National birth register (THL/974/5.05.00/2017).Breastmilk collection and HMO analysisMothers from the STEPS Study were asked to collect breastmilk samples when the infant was approximately 3 months old. In total, 812 of the 1797 mothers (45%) provided a breastmilk sample. There were only slight differences in maternal and child characteristics between the participants providing breastmilk samples and the total STEPS Study cohort40. Altogether, 795 breastmilk samples were included in this study (excluding the duplicate observations and the 2nd born twins, samples with technical unclarity or insufficient sample quantity, one breastmilk sample collected notably later than the other samples, at infant age of 14.5 months (range for the other breastmilk samples: 0.6–6.07 months), one sample with missing information on the date of collection and six mothers missing data on residential green environment) (Supplementary Fig. 2). Mothers received written instructions for the collection of breastmilk samples: samples were collected by manual expression in the morning from one single breast, first milking a few drops to waste before collecting the actual sample (~ 10 ml) into a plastic container (pre-feed sample). The samples were stored in the fridge and the mothers brought the samples to the research center or the samples were collected from their homes on the day of sampling. All samples were frozen and stored at − 70 °C until analysis.High Performance Liquid Chromatography (HPLC) was used to identify HMOs in breastmilk as previously described40,57,58 at the University of California, San Diego (methods described in detail in Berger et al.58). Milk samples were spiked with raffinose (a non-HMO carbohydrate) as an internal standard to allow absolute quantification. HMOs were extracted by high-throughput solid-phase extraction, fluorescently labelled, and measured using HPLC with fluorescent detection (HPLC-FLD)58. Absolute concentrations for the 19 HMOs were calculated based on standard retention times and corrected for internal standard recovery. Quantified HMOs included: 2′-fucosyllactose (2′FL), 3-fucosyllactose (3FL), lacto-N-neotetraose (LNnT), 3′-sialyllactose (3′SL), difucosyllactose (DFlac), 6′-sialyllactose (6′SL), lacto-N-tetraose (LNT), lacto-Nfucopentaose (LNFP) I, LNFP II, LNFP III, sialyl-LNT (LST) b, LSTc, difucosyllacto-LNT (DFLNT), lacto-N-hexaose (LNH), disialyllacto-N-tetraose (DSLNT), fucosyllacto-Nhexaose (FLNH), difucosyllacto-N-hexaose (DFLNH), fucodisialyllacto-lacto-N-hexaose (FDSLNH) and disialyllacto-N-hexaose (DSLNH). HMOs were also summed up to seven groups based on structural features: small HMOs (2′FL, 3FL, 3′SL, 6′SL, and DFLac), type 1 HMOs (LNT, LNFP I, LNFP II, LSTb, DSLNT), type 2 HMOs (LNnT, LNFP III, LSTc), α-1-2-fucosylated HMOs (2’FL, LNFP I), terminal α-2-6-sialylated HMOs (6′SL, LSTc), internal α-2-6-sialylated HMOs (DSLNT, LSTb), terminal α-2-3-sialylated HMOs (3′SL, DSLNT). The total concentration of HMOs was calculated as the sum of the 19 oligosaccharides. HMO-bound fucose and HMO-bound sialic acid were calculated on a molar basis. The proportion of each HMO comprising the total HMO concentration was also calculated. HMO Simpson’s diversity was calculated as Simpson’s Reciprocal Index 1/D, which is the reciprocal sum of the square of the relative abundance of each of the measured 19 HMOs57,59. The higher the diversity value, the more heterogenous is the HMO composition in the sample.Properties of the residential green environmentThe selected residential green environment variables measure the properties of the green environments surrounding the homes of the participants and do not include any measures of the house characteristics, indoor environment or the actual use of green spaces by the participants. The residential green environment variables were selected due to their previously observed associations with residential microbiota and health33,34,35. The variables of the residential green environments were derived from multispectral satellite images series, with a 30 m × 30 m of spatial resolution (Landsat TM 5, National Aeronautics and Space Administration—NASA) and land cover data (CORINE). We used Landsat TM images obtained over the summertime (June–August, greenest months in Finland), to minimize the seasonal variation of living vegetation and cloud cover as well as to better identify vegetation areas and maximise the contrast in our estimated exposure. In each selected Landsat TM 5 images, the cloud was masked out, and the Normalized Difference Vegetation Index (NDVI)36 was calculated. The final NDVI map was the mean of NDVI images collected over three consecutive years (2008–2010), to make an NDVI map with non-missing values due to cloud cover for the study area. NDVI map measures the vegetation cover, vitality and density. The NDVI can get values ranging from − 1 to 1 where values below zero represent water surfaces, values close to zero indicate areas with low intensity of living vegetation and values close to one indicate high abundance of living vegetation. For the analyses, areas covered by water were removed and the value ranged from 0 to 1, to prevent negative values for underestimating the greenness values of the residential area like in some prior studies60. We assumed that summertime NDVI identified the green space and vegetation density well, but greenness intensity might vary seasonally.Second, we used calculated indicators related to the diversity and naturalness of the land cover from CORINE Land Cover data sets of the year 201261. The 12 land cover types include: (1) Residential area, (2) Industrial/commercial area, (3) Transport network, (4) Sport/leisure, (5) Agriculture, (6) Broad-leave forest, (7) Coniferous forest, (8) Mixed forest, (9) Shrub/grassland, (10) Bare surface, (11) Wetland, and (12) Water bodies. From this information, we calculated two vegetation cover indexes. The Vegetation Cover Diversity Index (Simpson’s Diversity Index of Vegetation Cover, VCDI)37, only includes vegetation classes from CORINE land cover types (categories 5–9 and 11). VCDI approaches 1 as the number of different vegetation classes increases and the proportional distribution of area among the land use classes becomes more equitable. Furthermore, because we were particularly interested in the natural vegetation cover in the residential area, we calculated the area-weighted Naturalness Index (NI)38. This is an integrated indicator used to measure the human impact and degree of all human interventions on ecological components. The index is based on CORINE Land Cover data but reclassified to 15 classes. Residential areas have been divided to two classes: Continuous residential area (High density buildings) and Discontinuous residential area (Low density, mostly individual houses area). Agricultural area has also been divided to two classes: Agricultural area (Cropland) and Pasture as well as class 9 (Shrub/grassland) has been separated to Woodland and Natural grassland. Assignment of CORINE Land Cover classes to degrees of naturalness has been made based on Walz and Stein 201438. The area-weighted NI range from 1 to 7, where low values represent low human impact (≤ 3 = Natural), medium values moderate human impact (4–5 = Semi-natural) and high values strong human impact (6–7 = Non-Natural). To ease the interpretation of results and to correspond to the same direction than the other residential green environment variables, we have reverse-scaled the NI values, so that higher values illustrate more natural residential areas.Background factorsAs genetics is strongly linked to HMO composition, maternal secretor status was determined by high abundance (secretor) or near absence (non-secretor) of the HMO 2’FL in the breastmilk samples. Mothers with active secretor (Se) genes and FUT2 enzyme produce high amounts of α-1-2-fucosylated HMOs such as 2′-fucosyllactose (2′FL), whereas in the breastmilk of non-secretor mothers these HMOs are almost absent. Beyond genetics, other maternal and infant characteristics may influence HMO composition. So far, several associations have been reported, including lactation stage, maternal pre-pregnancy BMI, maternal age, parity, maternal diet, mode of delivery, infant gestational age and infant sex22,40. Information on the potential confounding factors, child sex, birth weight, maternal age at birth, number of previous births, marital status, maternal occupational status, smoking during pregnancy (before and during pregnancy), maternal pre-pregnancy BMI, mode of delivery, duration of pregnancy and maternal diseases [including both maternal disorders predominantly related to pregnancy such as pre-eclampsia and gestational diabetes and chronic diseases (diseases of the nervous, circulatory, respiratory, digestive, musculoskeletal and genitourinary systems, cancer and mental and behavioral disorders, according to ICD-10 codes, i.e. WHO International Classification of Diseases Tenth Revision)], were obtained from Medical Birth Registers. Self-administered questionnaires upon recruitment provided information on family net income and maternal education level. Those who had no professional training or a maximum of an intermediate level of vocational training (secondary education) were classified as “basic”. Those who had studied at a University of Applied Sciences or higher (tertiary education) were classified as “advanced”. The advanced level included any academic degree (bachelor’s, master’s, licentiate or doctoral degree). Maternal diet quality was assessed in late pregnancy using the Index of Diet Quality (IDQ62) which measures adherence to health promoting diet and nutrition recommendations. The IDQ score was used in its original form by setting the statistically defined cut-off value at 10, with scores below 10 points indicating unhealthy diets and non-adherence and scores of 10–15 points indicating a health-promoting diet and adherence dietary guidelines. Lactation time postpartum (child age) and season were received from the recorded breastmilk collection dates. Lactation status (exclusive/partial/unknown breastfeeding) at the time of breastmilk collection were gathered from follow-up diaries. From partially breastfeeding mothers (n = 277) 253 had started formula feeding and 28 solids at the time of milk collection. Last, a summary z score representing socio-economic disadvantage in the residential area was obtained from Statistics Finland grid database for the year 2009 and is based on the proportion of adults with low level of education, the unemployment rate, and proportion of people living in rented housing at each participant’s residential area55.Statistical analysesTo harmonize the residential green environment variables we calculated the mean values for NDVI, VCDI and NI in 750 × 750 m squares (and 250 × 250 m) around participant homes in a Geographical Information System (QGIS, www.qgis.org). The same grid sizes were used to calculate residential socioeconomic disadvantage in the residential area55 at the time of child birth. The geographical coordinates (latitude/longitude) of the cohort participants’ home address (795 mothers) were obtained from the Population Register Centre at the time of their child birth.The background characteristics of the mothers and children are given as means and standard deviations (SD) for continuous variables and percentages for categorical variables. Due to non-normal distribution, natural logarithmic transformation was performed for all HMO variables (19 individual components, sum of HMOs, HMO-bound sialic acid, HMO-bound fucose and HMO groups (all in nmol/mL)) except for HMO diversity. Associations between each background factor and HMO diversity and 19 individual HMO components were analysed with univariate generalized linear models to identify factors independently associated with HMO composition. All factors demonstrating a significant association (p  More

Sample collection and postmortem evaluationBald eagle carcasses, and/or oropharyngeal and cloacal swabs were collected in the field and submitted to the Southeastern Cooperative Wildlife Disease Study Research and Diagnostic Service. In some cases, live bald eagles were found moribund and transported to wildlife rehabilitation clinics and either died in transit or soon after arrival. Carcasses underwent postmortem evaluation, including gross and histopathology. Tissue samples [heart, brain, kidney, spleen, lung, adrenal gland, pancreas, liver, small and large intestine, and cloacal bursa (if present)] were fixed in 10% neutral buffered formalin and routinely processed for histopathology23 at the Athens Veterinary Diagnostic Laboratory. Histopathology was assessed by a board-certified veterinary pathologist.Additional bald eagle and waterfowl species mortality dataData on wild bird deaths attributed to highly pathogenic influenza A viruses were retrieved from the U.S. Department of Agriculture, Animal and Plant Health Inspection Service website, at: https://www.aphis.usda.gov/aphis/ourfocus/animalhealth/animal-disease-information/avian/avian-influenza/hpai-2022/2022-hpai-wild-birds. These data are publicly available and include state, county, date detected, and species of individual birds that tested positive for HP IAV.ImmunohistochemistryImmunohistochemistry (IHC) for avian influenza virus was performed in select cases on brain, pancreas, spleen, liver, and/or adrenal gland at the Athens Veterinary Diagnostic Laboratory. IHC was performed on an automated stainer (Nemesis 3600, Biocare Medical). Polyclonal antiserum against influenza A virus was used as the primary antibody (ab155877, Abcam), diluted 1:3000, and incubated for 60 min at 37 °C with agent-positive control. Antigen retrieval was with Target Retrieval Solution (S2367, Dako) pH (10x) at 110 °C for 15 min. Enzyme blockage was via 3% H2O2 for 20 min (H324-500, Fisher Scientific); protein blockage was with Universal Blocking Reagent (10x) Power Block diluted at 1:10 for 5 min (HK085-5 K, BioGenex); link was by biotinylated rabbit anti-goat (BA-5000, Vector) at a 1:100 dilution for 10 min with 4 + streptavidin alkaline phosphatase label for 10 min (AP605H, BioCare Medical). Staining was with warp red chromogen kit for 5 min (WR8065, BioCare Medical). Known influenza A-virus positive control tissues were tested alongside each case.Polymerase chain reactionOropharyngeal and cloacal swabs from bald eagle carcasses were pooled for each individual eagle and tested by real-time reverse transcription polymerase chain reaction (rRT-PCR). Briefly, swabs samples were extracted with the KingFisher magnetic particle processer using the MagMAX-96 AI/ND Viral RNA isolation Kit (Ambion/Applied Biosystems, Foster City, CA) following a modified MagMAX-S protocol24. Resultant nucleic acids were screened against primers specific for H5 IAV in rRT-PCR; samples that yielded a cycle threshold value  More

De Goeij, J. M. et al. Surviving in a marine desert: The sponge loop retains resources within coral reefs. Science 80(342), 108–110 (2013).Article 

Google Scholar 
Graham, N. A. J. & Nash, K. L. The importance of structural complexity in coral reef ecosystems. Coral Reefs 32, 315–326 (2013).Article 
ADS 

Google Scholar 
Perry, C. T. et al. Loss of coral reef growth capacity to track future increases in sea level. Nature 558, 396–400 (2018).Article 
ADS 

Google Scholar 
Enochs, I. C. & Manzello, D. P. Responses of cryptofaunal species richness and trophic potential to coral reef habitat degradation. Diversity 4, 94–104 (2012).Article 

Google Scholar 
Newman, S. P. et al. Reef flattening effects on total richness and species responses in the Caribbean. J. Anim. Ecol. 84, 1678–1689 (2015).Article 

Google Scholar 
Ferrario, F. et al. The effectiveness of coral reefs for coastal hazard risk reduction and adaptation. Nat. Commun. 5, 1–9 (2014).Article 

Google Scholar 
Storlazzi, C. D. et al. Rigorously valuing the potential coastal hazard risk reduction provided by coral reef restoration in Florida and Puerto Rico. U.S. Geol. Surv. 2, 1–24 (2021).
Google Scholar 
Cornwall, C. E. et al. Global declines in coral reef calcium carbonate production under ocean acidification and warming. Proc. Natl. Acad. Sci. U. S. A. https://doi.org/10.1073/pnas.2015265118 (2021).Article 

Google Scholar 
Ries, J. B., Cohen, A. L. & McCorkle, D. C. Marine calcifiers exhibit mixed responses to CO2-induced ocean acidification. Geology 37, 1131–1134 (2009).Article 
ADS 

Google Scholar 
Achlatis, M. et al. Sponge bioerosion on changing reefs: Ocean warming poses physiological constraints to the success of a photosymbiotic excavating sponge. Sci. Rep. 7, 1–13 (2017).Article 

Google Scholar 
Perry, C. T. & Alvarez-Filip, L. Changing geo-ecological functions of coral reefs in the Anthropocene. Funct. Ecol. 33, 976–988 (2019).
Google Scholar 
Manzello, D. P., Enochs, I. C., Kolodziej, G. & Carlton, R. Recent decade of growth and calcification of Orbicella faveolata in the Florida Keys: An inshore-offshore comparison. Mar. Ecol. Prog. Ser. 521, 81–89 (2015).Article 
ADS 

Google Scholar 
de Bakker, D. M., van Duyl, F. C., Perry, C. T. & Meesters, E. H. Extreme spatial heterogeneity in carbonate accretion potential on a Caribbean fringing reef linked to local human disturbance gradients. Glob. Chang. Biol. 25, 4092–4104 (2019).Article 
ADS 

Google Scholar 
Wisshak, M., Schönberg, C. H. L., Form, A. & Freiwald, A. Ocean acidification accelerates reef bioerosion. PLoS ONE 7, e45124 (2012).Article 
ADS 

Google Scholar 
Webb, A. E. et al. Combined effects of experimental acidification and eutrophication on reef sponge bioerosion rates. Front. Mar. Sci. 4, 311 (2017).Article 

Google Scholar 
Perry, C. T. et al. Estimating rates of biologically driven coral reef framework production and erosion: A new census-based carbonate budget methodology and applications to the reefs of Bonaire. Coral Reefs 31, 853–868 (2012).Article 
ADS 

Google Scholar 
Molina-Hernández, A., González-Barrios, F. J., Perry, C. T. & Álvarez-Filip, L. Two decades of carbonate budget change on shifted coral reef assemblages: Are these reefs being locked into low net budget states?: Caribbean reefs carbonate budget trends. Proc. R. Soc. B Biol. Sci. 287, 20202305 (2020).Article 

Google Scholar 
Toth, L. T., Courtney, T. A., Colella, M. A., Kupfner, S. A. & Robert, J. The past, present, and future of coral reef growth in the Florida Keys. Glob. Change Biol. 28(17), 5294–5309. https://doi.org/10.1111/gcb.16295 (2022).Article 

Google Scholar 
Perry, C. T. et al. Caribbean-wide decline in carbonate production threatens coral reef growth. Nat. Commun. 4, 1402–1407 (2013).Article 
ADS 

Google Scholar 
Enochs, I. C. et al. Ocean acidification enhances the bioerosion of a common coral reef sponge: Implications for the persistence of the Florida Reef Tract. Bull. Mar. Sci. 91, 271–290 (2015).Article 

Google Scholar 
Bellwood, D. R. et al. Coral reef conservation in the Anthropocene: Confronting spatial mismatches and prioritizing functions. Biol. Conserv. 236, 604–615 (2019).Article 

Google Scholar 
van Hooidonk, R., Maynard, J. A., Liu, Y. & Lee, S. K. Downscaled projections of Caribbean coral bleaching that can inform conservation planning. Glob. Chang. Biol. 21, 3389–3401 (2015).Article 
ADS 

Google Scholar 
Pandolfi, J. M., Connolly, S. R., Marshall, D. J. & Cohen, A. L. Projecting coral reef futures under global warming and ocean acidification. Science 80(333), 418 (2011).Article 
ADS 

Google Scholar 
Teneva, L. et al. Predicting coral bleaching hotspots: The role of regional variability in thermal stress and potential adaptation rates. Coral Reefs 31, 1–12 (2012).Article 
ADS 

Google Scholar 
Albright, R., Langdon, C. & Anthony, K. R. N. Dynamics of seawater carbonate chemistry, production, and calcification of a coral reef flat, central great Barrier Reef. Biogeosciences 10, 6747–6758 (2013).Article 
ADS 

Google Scholar 
Van Hooidonk, R., Maynard, J. A., Manzello, D. & Planes, S. Opposite latitudinal gradients in projected ocean acidification and bleaching impacts on coral reefs. Glob. Chang. Biol. 20, 103–112 (2014).Article 
ADS 

Google Scholar 
Van Hooidonk, R. et al. Local-scale projections of coral reef futures and implications of the paris agreement. Sci. Rep. 6, 1–8 (2016).
Google Scholar 
Lee, J.-Y. et al. (2021) Future global climate: Scenario-based projections and near-term information supplementary material climate change 2021: the physical science basis. Contribution of working group i to the sixth assessment report of the intergovernmental panel on climate change doi:https://doi.org/10.1017/9781009157896.006McCulloch, M., Falter, J., Trotter, J. & Montagna, P. Coral resilience to ocean acidification and global warming through pH up-regulation. Nat. Clim. Chang. 2, 623–627 (2012).Article 
ADS 

Google Scholar 
Okazaki, R. R. et al. Species-specific responses to climate change and community composition determine future calcification rates of Florida Keys reefs. Glob. Chang. Biol. 23, 1023–1035 (2017).Article 
ADS 

Google Scholar 
Kornder, N. A., Riegl, B. M. & Figueiredo, J. Thresholds and drivers of coral calcification responses to climate change. Glob. Chang. Biol. 24, 5084–5095 (2018).Article 
ADS 

Google Scholar 
Van Hooidonk, R., Maynard, J. A. & Planes, S. Temporary refugia for coral reefs in a warming world. Nat. Clim. Chang. 3, 508–511 (2013).Article 
ADS 

Google Scholar 
Logan, C. A., Dunne, J. P., Eakin, C. M. & Donner, S. D. Incorporating adaptive responses into future projections of coral bleaching. Glob. Chang. Biol. 20, 125–139 (2014).Article 
ADS 

Google Scholar 
Kennedy, E. V. et al. Avoiding coral reef functional collapse requires local and global action. Curr. Biol. 23, 912–918 (2013).Article 

Google Scholar 
Ruzicka, R. R. et al. Temporal changes in benthic assemblages on Florida Keys reefs 11 years after the 1997/1998 El Niño. Mar. Ecol. Prog. Ser. 489, 125–141 (2013).Article 
ADS 

Google Scholar 
Manzello, D. P., Enochs, I. C., Kolodziej, G., Carlton, R. & Valentino, L. Resilience in carbonate production despite three coral bleaching events in 5 years on an inshore patch reef in the Florida Keys. Mar. Biol. 165, 1–11 (2018).Article 

Google Scholar 
Gintert, B. E. et al. Marked annual coral bleaching resilience of an inshore patch reef in the Florida Keys: A nugget of hope, aberrance, or last man standing?. Coral Reefs 37, 533–547 (2018).Article 
ADS 

Google Scholar 
Maynard, J. A., Anthony, K. R. N., Marshall, P. A. & Masiri, I. Major bleaching events can lead to increased thermal tolerance in corals. Mar. Biol. 155, 173–182 (2008).Article 

Google Scholar 
Sampayo, E. M., Ridgway, T., Bongaerts, P. & Hoegh-Guldberg, O. Bleaching susceptibility and mortality of corals are determined by fine-scale differences in symbiont type. Proc. Natl. Acad. Sci. U. S. A. 105, 10444–10449 (2008).Article 
ADS 

Google Scholar 
Silverstein, R. N., Correa, A. M. S. & Baker, A. C. Specificity is rarely absolute in coral–algal symbiosis: Implications for coral response to climate change. Proc. R. Soc. B Biol. Sci. 279, 2609–2618 (2012).Article 

Google Scholar 
Barshis, D. J. et al. Genomic basis for coral resilience to climate change. Proc. Natl. Acad. Sci. U. S. A. 110, 1387–1392 (2013).Article 
ADS 

Google Scholar 
Voolstra, C. R. et al. Rapid evolution of coral proteins responsible for interaction with the environment. PLoS ONE 6(5), e20392 (2011).Article 
ADS 

Google Scholar 
Hughes, T. P. et al. Global warming transforms coral reef assemblages. Nature 556, 492–496 (2018).Article 
ADS 

Google Scholar 
Palumbi, S. R., Barshis, D. J., Traylor-Knowles, N. & Bay, R. A. Mechanisms of reef coral resistance to future climate change. Science 80(344), 895 (2014).Article 
ADS 

Google Scholar 
Tribollet, A., Chauvin, A. & Cuet, P. Carbonate dissolution by reef microbial borers: A biogeological process producing alkalinity under different pCO 2 conditions. Facies 65, 1–10 (2019).Article 

Google Scholar 
Chaves-Fonnegra, A. et al. Bleaching events regulate shifts from corals to excavating sponges in algae-dominated reefs. Glob. Chang. Biol. 24, 773–785 (2018).Article 
ADS 

Google Scholar 
Enochs, I. C. et al. Upwelling and the persistence of coral-reef frameworks in the eastern tropical Pacific. Ecol. Monogr. 91, 1–16 (2021).Article 

Google Scholar 
Van Westen, R. M. & Dijkstra, H. A. Ocean eddies strongly affect global mean sea-level projections. Sci. Adv. 7, 1–12 (2021).
Google Scholar 
DeMerlis, A. et al. Pre-exposure to a variable temperature treatment improves the response of Acropora cervicornis to acute thermal stress. Coral Reefs https://doi.org/10.1007/s00338-022-02232-z (2022).Article 

Google Scholar 
Webb, A. E. et al. Quantifying functional consequences of habitat degradation on a Caribbean coral reef. Biogeosciences 18, 6501–6516 (2021).Article 
ADS 

Google Scholar 
Silbiger, N. J. et al. Nutrient pollution disrupts key ecosystem functions on coral reefs. Proc. R. Soc. B Biol. Sci. 285, 2–10 (2018).
Google Scholar 
DeCarlo, T. M. et al. Coral macrobioerosion is accelerated by ocean acidification and nutrients. Geology 43, 7–10 (2015).Article 
ADS 

Google Scholar 
Wooldridge, S. A. Water quality and coral bleaching thresholds: Formalising the linkage for the inshore reefs of the Great Barrier Reef. Australia. Mar. Pollut. Bull. 58, 745–751 (2009).Article 

Google Scholar 
Eyring, V. et al. Overview of the coupled model intercomparison project phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev. https://doi.org/10.5194/gmd-9-1937-2016 (2016).O’Neill, B. C., Kriegler, E., Riahi, K. & Ebi, K. L. A new scenario framework for climate change research : The concept of shared socioeconomic pathways. Clim. Change https://doi.org/10.1007/s10584-013-0905-2 (2014).Article 

Google Scholar 
O’Neill, B. C. et al. The scenario model intercomparison project ( ScenarioMIP ) for CMIP6. Geosci. Model Dev. 9(9), 3461–3482 (2018).Article 
ADS 

Google Scholar 
Towle, E. K. et al. (2021) National coral reef monitoring plan.Kuffner, I. B., Hickey, T. D. & Morrison, J. M. Calcification rates of the massive coral Siderastrea siderea and crustose coralline algae along the Florida Keys (USA) outer-reef tract. Coral Reefs 32, 987–997 (2013).Article 
ADS 

Google Scholar 
Manzello, D. P., Enochs, I. C., Kolodziej, G. & Carlton, R. Coral growth patterns of Montastraea cavernosa and Porites astreoides in the Florida Keys: The importance of thermal stress and inimical waters. J. Exp. Mar. Bio. Ecol. 471, 198–207 (2015).Article 

Google Scholar 
Gattuso, J.P. et al. (2015) Package ‘ seacarb ’Donner, S. D., Skirving, W. J., Little, C. M., Oppenheimer, M. & Hoegh-Gulberg, O. Global assessment of coral bleaching and required rates of adaptation under climate change. Glob. Chang. Biol. 11, 2251–2265 (2005).Article 
ADS 

Google Scholar 
van Hooidonk, R. & Huber, M. Quantifying the quality of coral bleaching predictions. Coral Reefs 28, 579–587 (2009).Article 
ADS 

Google Scholar 
Yee, S. H. & Barron, M. G. Predicting coral bleaching in response to environmental stressors using 8 years of global-scale data. Environ. Monit. Assess. 161, 423–438 (2010).Article 

Google Scholar 
Liu, G., Strong, A. E. & Skirving, W. Remote sensing of sea surface temperatures during 2002 barrier reef coral bleaching (Eos, Washington, DC, 2003).Book 

Google Scholar 
van Hooidonk, R. et al. Projections of future coral bleaching conditions using IPCC CMIP6 models: Climate policy implications, management applications, and Regional Seas summaries. (2020).Kinsey, D. W. & Hopley, D. The significance of coral reefs as global carbon sinks-response to Greenhouse. Glob. Planet. Change 3, 363–377 (1991).Article 
ADS 

Google Scholar 
van Westen, R. M. et al. Ocean model resolution dependence of Caribbean sea-level projections. Sci. Rep. 10, 1–11 (2020).
Google Scholar  More

FAO. Handbook on crop statistics: improving methods for measuring crop area, production and yield. (FAO, Rome, Italy, 2018).FAO. Land use statistics and indicators: global, regional and county trends 1990-2019. FAOSTAT Anal. Brief Ser. No 28 (2021).Potapov, P. et al. Global maps of cropland extent and change show accelerated cropland expansion in the twenty-first century. Nat. Food 1–10 (2021) https://doi.org/10.1038/s43016-021-00429-zFAO. A system of integrated agricultural censuses and surveys. (FAO, 2005).FAO. Land use statistics and indicators. Global, regional and country trends, 2000–2020. (FAO, Rome, Italy, 2022).Loveland, T. R. et al. Development of a global land cover characteristics database and IGBP DISCover from 1 km AVHRR data. Int. J. Remote Sens. 21, 1303–1330 (2000).Article 

Google Scholar 
Zanaga, D. et al. ESA WorldCover 10 m 2020 v100. (2021) https://doi.org/10.5281/zenodo.5571936Cochran, W. G. Sampling techniques. (Wiley, 1977).Stehman, S. V. Estimating area and map accuracy for stratified random sampling when the strata are different from the map classes. Int. J. Remote Sens. 35, 4923–4939 (2014).Article 

Google Scholar 
Tsujino, R., Kaijisa, T. & Yumoto, T. Causes and history of forest loss in Cambodia. Int. For. Rev. 21, 372–384 (2019).
Google Scholar 
Hu, Q. et al. Global cropland intensification surpassed expansion between 2000 and 2010: A spatio-temporal analysis based on GlobeLand30. Sci. Total Environ. 746, 141035 (2020).Grainger, A. Difficulties in tracking the long-term global trend in tropical forest area. Proc. Natl Acad. Sci. 105, 818–823 (2008).Article 
ADS 
CAS 

Google Scholar 
FAO. FAOSTAT. https://www.fao.org/faostat/en/#home (2021). More