Philippa Kaur

Buisson, E., Archibald, S., Fidelis, A. & Suding, K. N. Ancient grasslands guide ambitious goals in grassland restoration. Science 377, 594–598. https://doi.org/10.1126/science.abo4605 (2022).Article 
PubMed 

Google Scholar 
Archibald, S. et al. Biological and geophysical feedbacks with fire in the Earth system. Environ. Res. Lett. 13, 033003. https://doi.org/10.1088/1748-9326/aa9ead (2018).Article 

Google Scholar 
Keeley, J. E., Pausas, J. G., Rundel, P. W., Bond, W. J. & Bradstock, R. A. Fire as an evolutionary pressure shaping plant traits. Trends Plant Sci. 16, 406–411. https://doi.org/10.1016/j.tplants.2011.04.002 (2011).Article 
PubMed 

Google Scholar 
Whitman, T. et al. Soil bacterial and fungal response to wildfires in the Canadian boreal forest across a burn severity gradient. Soil Biol. Biochem. 138, 107571. https://doi.org/10.1016/j.soilbio.2019.107571 (2019).Article 

Google Scholar 
Platt, W. J., Ellair, D. P., Huffman, J. M., Potts, S. E. & Beckage, B. Pyrogenic fuels produced by savanna trees can engineer humid savannas. Ecol. Monogr. 86, 352–372. https://doi.org/10.1002/ecm.1224 (2016).Article 

Google Scholar 
He, T., Lamont, B. B. & Pausas, J. G. Fire as a key driver of Earth’s biodiversity. Biol. Rev. 94, 1983–2010. https://doi.org/10.1111/brv.12544 (2019).Article 
PubMed 

Google Scholar 
Bond, W. J. & Keeley, J. E. Fire as a global ‘herbivore’: the ecology and evolution of flammable ecosystems. Trends Ecol. Evol. 20, 387–394 (2005).Article 
PubMed 

Google Scholar 
Hopkins, J. R., Huffman, J. M., Platt, W. J. & Sikes, B. A. Frequent fire slows microbial decomposition of newly deposited fine fuels in a pyrophilic ecosystem. Oecologia 193, 631–643. https://doi.org/10.1007/s00442-020-04699-5 (2020).Article 
PubMed 

Google Scholar 
Platt, W. J., Orzell, S. L. & Slocum, M. G. Seasonality of fire weather strongly influences fire regimes in south Florida savanna-grassland landscapes. PLoS ONE 10, e0116952 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Semenova-Nelsen, T. A., Platt, W. J., Patterson, T. R., Huffman, J. & Sikes, B. A. Frequent fire reorganizes fungal communities and slows decomposition across a heterogeneous pine savanna landscape. New Phytol. 224, 916–927. https://doi.org/10.1111/nph.16096 (2019).Article 
PubMed 

Google Scholar 
Hansen, P. M., Semenova-Nelsen, T. A., Platt, W. J. & Sikes, B. A. Recurrent fires do not affect the abundance of soil fungi in a frequently burned pine savanna. Fungal Ecol. 42, 100852. https://doi.org/10.1016/j.funeco.2019.07.006 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Köster, K. et al. The long-term impact of low-intensity surface fires on litter decomposition and enzyme activities in boreal coniferous forests. Int. J. Wildland Fire 25, 618–618 (2016).Article 

Google Scholar 
Beals, K. K., Scearce, A. E., Swystun, A. T. & Schweitzer, J. A. Belowground mechanisms for oak regeneration: Interactions among fire, soil microbes, and plant community alter oak seedling growth. For. Ecol. Manage. 503, 119774. https://doi.org/10.1016/j.foreco.2021.119774 (2022).Article 

Google Scholar 
Huffman, M. S. & Madritch, M. D. Soil microbial response following wildfires in thermic oak-pine forests. Biol. Fertil. Soils 54, 985–997 (2018).Article 

Google Scholar 
Certini, G. Effects of fire on properties of forest soils: A review. Oecologia 143, 1–10 (2005).Article 
PubMed 

Google Scholar 
Allison, S. D. & Martiny, J. B. H. Resistance, resilience, and redundancy in microbial communities. Proc. Natl. Acad. Sci. 105, 11512–11519. https://doi.org/10.1073/pnas.0801925105 (2008).Article 
PubMed 
PubMed Central 

Google Scholar 
Badía, D. et al. Burn effects on soil properties associated to heat transfer under contrasting moisture content. Sci. Total Environ. 601–602, 1119–1128. https://doi.org/10.1016/j.scitotenv.2017.05.254 (2017).Article 
PubMed 

Google Scholar 
Mino, L. et al. Watershed and fire severity are stronger determinants of soil chemistry and microbiomes than within-watershed woody encroachment in a tallgrass prairie system. FEMS Microbiol. Ecol. 97, fiab154. https://doi.org/10.1093/femsec/fiab154 (2021).Article 
PubMed 

Google Scholar 
Mataix-Solera, J., García-Orenes, F., Bárcenas-Moreno, G. & Torres, M. Forest Fire Effects on Soil Microbiology.
In Fire Effects on Soils and Restoration Strategies, (eds A. Cerdà & P. Robichaud) 133–175 (Science Publishers, Inc., 2009). https://doi.org/10.1201/9781439843338-c5.McLauchlan, K. K. et al. Fire as a fundamental ecological process: Research advances and frontiers. J. Ecol. 108, 2047–2069. https://doi.org/10.1111/1365-2745.13403 (2020).Article 

Google Scholar 
Prober, S. M. et al. Plant diversity predicts beta but not alpha diversity of soil microbes across grasslands worldwide. Ecol. Lett. 18, 85–95 (2015).Article 
PubMed 

Google Scholar 
Veldman, J. W. et al. Toward an old-growth concept for grasslands, savannas, and woodlands. Front. Ecol. Environ. 13, 154–162. https://doi.org/10.1890/140270 (2015).Article 

Google Scholar 
Peet, R., Platt, W. & Costanza, J. Fire-maintained Pine Savannas and Woodlands of the Southeastern US Coastal Plain. in Ecology and Recovery of Eastern Old-Growth Forests (eds Barton, A. M. & Keeton, W. S.) Ch. 3, (2018).Costanza, J. K., Terando, A. J., McKerrow, A. J. & Collazo, J. A. Modeling climate change, urbanization, and fire effects on Pinus palustris ecosystems of the southeastern US. J. Environ. Manage. 151, 186–199. https://doi.org/10.1016/j.jenvman.2014.12.032 (2015).Article 
PubMed 

Google Scholar 
Ibanez, T. et al. Altered cyclone–fire interactions are changing ecosystems. Trends Plant Sci. https://doi.org/10.1016/j.tplants.2022.08.005 (2022).Article 
PubMed 

Google Scholar 
Robertson, K. M., Platt, W. J. & Faires, C. E. Patchy fires promote regeneration of longleaf pine (Pinus palustris Mill.) in pine savannas. Forests https://doi.org/10.3390/f10050367 (2019).Article 

Google Scholar 
Platt, W. J., Evans, G. W. & Rathbun, S. L. The population dynamics of a long-lived conifer (Pinus palustris). Am. Nat. 131, 491–525 (1988).Article 

Google Scholar 
Noel, J., Platt, W. J. & Moser, E. Structural characteristics of old- and second-growth stands of longleaf pine (Pinus palustris) in the gulf coastal region of the USA. Conserv. Biol. 12, 533–548. https://doi.org/10.1111/j.1523-1739.1998.96124.x (1998).Article 

Google Scholar 
Ellair, D. P. & Platt, W. J. Fuel composition influences fire characteristics and understorey hardwoods in pine savanna. J. Ecol. 101, 192–201. https://doi.org/10.1111/1365-2745.12008 (2013).Article 

Google Scholar 
Senn, S. et al. The functional biogeography of eDNA metacommunities in the post-fire landscape of the Angeles national forest. Microorganisms https://doi.org/10.3390/microorganisms10061218 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Ammitzboll, H., Jordan, G. J., Baker, S. C., Freeman, J. & Bissett, A. Contrasting successional responses of soil bacteria and fungi to post-logging burn severity. For. Ecol. Manage. 508, 120059. https://doi.org/10.1016/j.foreco.2022.120059 (2022).Article 

Google Scholar 
Rother, M. T., Huffman, J. M., Guiterman, C. H., Robertson, K. M. & Jones, N. A history of recurrent, low-severity fire without fire exclusion in southeastern pine savannas, USA. For. Ecol. Manage. 475, 118406. https://doi.org/10.1016/j.foreco.2020.118406 (2020).Article 

Google Scholar 
Noss, R. F. et al. How global biodiversity hotspots may go unrecognized: lessons from the North American Coastal Plain. Divers. Distrib. 21, 236–244. https://doi.org/10.1111/ddi.12278 (2015).Article 

Google Scholar 
Platt, W. J. Southeastern pine savannas. in Savannas, Barrens, and Rock Outcrop Plant Communities of North America, 23–51 (1999).Fill, J. M., Platt, W. J., Welch, S. M., Waldron, J. L. & Mousseau, T. A. Updating models for restoration and management of fiery ecosystems. For. Ecol. Manage. 356, 54–63 (2015).Article 

Google Scholar 
Fill, J. M., Davis, C. N. & Crandall, R. M. Climate change lengthens southeastern USA lightning-ignited fire seasons. Glob. Change Biol. 25, 3562–3569. https://doi.org/10.1111/gcb.14727 (2019).Article 

Google Scholar 
Weakley, A. Flora of the Southern and Mid-Atlantic States, (2015).Multivariate analysis of Ecological Data, Version 6.0 for Windows (MjM Software, Gleneden Beach, Oregon, USA, 2011).Caporaso, J. G. et al. Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proc. Natl. Acad. Sci. 108, 4516. https://doi.org/10.1073/pnas.1000080107 (2011).Article 
PubMed 

Google Scholar 
Renaud, G., Stenzel, U., Maricic, T., Wiebe, V. & Kelso, J. deML: Robust demultiplexing of Illumina sequences using a likelihood-based approach. Bioinformatics 31, 770–772. https://doi.org/10.1093/bioinformatics/btu719 (2015).Article 
PubMed 

Google Scholar 
Callahan, B. J. et al. DADA2: High-resolution sample inference from Illumina amplicon data. Nat. Methods 13, 581–583 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Xia, Z. et al. Conventional versus real-time quantitative PCR for rare species detection. Ecol. Evol. 8, 11799–11807. https://doi.org/10.1002/ece3.4636 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
Quast, C. et al. in Encyclopedia of Metagenomics: Genes, Genomes and Metagenomes: Basics, Methods, Databases and Tools (ed Nelson, K. E.) 626–635 (Springer US, 2015).Wang, Q., Garrity, G. M., Tiedje, J. M. & Cole, J. R. Naïve Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261. https://doi.org/10.1128/AEM.00062-07 (2007).Article 
PubMed 
PubMed Central 

Google Scholar 
Balvočiūtė, M. & Huson, D. H. SILVA, RDP, Greengenes, NCBI and OTT—how do these taxonomies compare?. BMC Genomics 18, 114. https://doi.org/10.1186/s12864-017-3501-4 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 1–21 (2014).Article 

Google Scholar 
Legendre, P. & Legendre, L. Numerical Ecology. (Elsevier, 2012).Jorgensen, B. Exponential dispersion models. J. R. Stat. Soc. Ser. B (Methodol.) 49, 127–162 (1987).MathSciNet 
MATH 

Google Scholar 
Team, R. C. R: A Language and Environment for Statistical Computing (2019).Wood, S. Package ‘mgcv’. R Package Version 1, 29 (2015).
Google Scholar 
Jones, G. M. & Tingley, M. W. Pyrodiversity and biodiversity: A history, synthesis, and outlook. Divers. Distrib. 28, 386–403. https://doi.org/10.1111/ddi.13280 (2022).Article 

Google Scholar 
Pfeiffer, B. et al. Leaf litter is the main driver for changes in bacterial community structures in the rhizosphere of ash and beech. Appl. Soil. Ecol. 72, 150–160 (2013).Article 

Google Scholar 
Purahong, W. et al. Life in leaf litter: novel insights into community dynamics of bacteria and fungi during litter decomposition. Mol. Ecol. https://doi.org/10.1111/mec.13739 (2016).Article 
PubMed 

Google Scholar 
Angst, Š et al. Tree species identity alters decomposition of understory litter and associated microbial communities: A case study. Biol. Fertil. Soils 55, 525–538. https://doi.org/10.1007/s00374-019-01360-z (2019).Article 

Google Scholar 
Liang, X., Yuan, J., Yang, E. & Meng, J. Responses of soil organic carbon decomposition and microbial community to the addition of plant residues with different C:N ratio. Eur. J. Soil Biol. 82, 50–55. https://doi.org/10.1016/j.ejsobi.2017.08.005 (2017).Article 

Google Scholar 
Bonanomi, G. et al. Litter chemistry explains contrasting feeding preferences of bacteria, fungi, and higher plants. Sci. Rep. 7, 9208. https://doi.org/10.1038/s41598-017-09145-w (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Li, W., Niu, S., Liu, X. & Wang, J. Short-term response of the soil bacterial community to differing wildfire severity in Pinus tabulaeformis stands. Sci. Rep. 9, 1148. https://doi.org/10.1038/s41598-019-38541-7 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Ficken, C. D. & Wright, J. P. Effects of fire frequency on litter decomposition as mediated by changes to litter chemistry and soil environmental conditions. PLoS ONE 12, e0186292 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Bani, A. et al. The role of microbial community in the decomposition of leaf litter and deadwood. Appl. Soil. Ecol. 126, 75–84. https://doi.org/10.1016/j.apsoil.2018.02.017 (2018).Article 

Google Scholar 
Bowd, E. J. et al. Direct and indirect effects of fire on microbial communities in a pyrodiverse dry-sclerophyll forest. J. Ecol. https://doi.org/10.1111/1365-2745.13903 (2022).Article 

Google Scholar 
Hart, S., Deluca, T., Newman, G., Mackenzie, M. D. & Boyle, S. Post-fire vegetative dynamics as drivers of microbial community structure and function in forest soils. For. Ecol. Manage. https://doi.org/10.1016/j.foreco.2005.08.012 (2005).Article 

Google Scholar 
López-Mondéjar, R. et al. Decomposer food web in a deciduous forest shows high share of generalist microorganisms and importance of microbial biomass recycling. ISME J. https://doi.org/10.1038/s41396-018-0084-2 (2018).Article 
PubMed 
PubMed Central 

Google Scholar 
Pérez-Valera, E., Verdú, M., Navarro Cano, J. & Goberna, M. Resilience to fire of phylogenetic diversity across biological domains. Mol. Ecol. https://doi.org/10.1111/mec.14729 (2018).Article 
PubMed 

Google Scholar 
Zhao, F. et al. Change in soil bacterial community during secondary succession depend on plant and soil characteristics. CATENA 173, 246–252 (2019).Article 

Google Scholar 
Mikita-Barbato, R. A., Kelly, J. J. & Tate, R. L. Wildfire effects on the properties and microbial community structure of organic horizon soils in the New Jersey Pinelands. Soil Biol. Biochem. 86, 67–76. https://doi.org/10.1016/j.soilbio.2015.03.021 (2015).Article 

Google Scholar 
Adkins, J., Docherty, K. M., Gutknecht, J. L. M. & Miesel, J. R. How do soil microbial communities respond to fire in the intermediate term? Investigating direct and indirect effects associated with fire occurrence and burn severity. Sci. Total Environ. 745, 140957. https://doi.org/10.1016/j.scitotenv.2020.140957 (2020).Article 
PubMed 

Google Scholar 
Ponder, F. Jr., Tadros, M. & Loewenstein, E. F. Microbial properties and litter and soil nutrients after two prescribed fires in developing savannas in an upland Missouri Ozark Forest. For. Ecol. Manage. 257, 755–763 (2009).Article 

Google Scholar 
Gołębiewski, M. et al. Rapid microbial community changes during initial stages of pine litter decomposition. Microb. Ecol. 77, 56–75. https://doi.org/10.1007/s00248-018-1209-x (2019).Article 
PubMed 

Google Scholar 
Coetsee, C., Bond, W. J. & February, E. C. Frequent fire affects soil nitrogen and carbon in an African savanna by changing woody cover. Oecologia 162, 1027–1034 (2010).Article 
PubMed 

Google Scholar 
Alcañiz, M., Outeiro, L., Francos, M. & Ubeda, X. Effects of prescribed fires on soil properties: A review. Sci Total Environ 613–614, 944–957. https://doi.org/10.1016/j.scitotenv.2017.09.144 (2018).Article 
PubMed 

Google Scholar 
Ferrenberg, S. et al. Changes in assembly processes in soil bacterial communities following a wildfire disturbance. ISME J. 7, 1102–1111 (2013).Article 
PubMed 
PubMed Central 

Google Scholar 
Kranz, C. & Whitman, T. Surface charring from prescribed burning has minimal effects on soil bacterial community composition two weeks post-fire in jack pine barrens. Appl. Soil. Ecol. 144, 134–138 (2019).Article 

Google Scholar 
Whitman, T., Woolet, J., Sikora, M., Johnson, D. B. & Whitman, E. Resilience in soil bacterial communities of the boreal forest from one to five years after wildfire across a severity gradient. Soil Biol. Biochem. 172, 108755. https://doi.org/10.1016/j.soilbio.2022.108755 (2022).Article 

Google Scholar 
Ammitzboll, H., Jordan, G. J., Baker, S. C., Freeman, J. & Bissett, A. Diversity and abundance of soil microbial communities decline, and community compositions change with severity of post-logging fire. Mol. Ecol. 30, 2434–2448. https://doi.org/10.1111/mec.15900 (2021).Article 
PubMed 

Google Scholar 
Maquia, I. S. A. et al. The nexus between fire and soil bacterial diversity in the African miombo woodlands of niassa special reserve, Mozambique. Microorganisms https://doi.org/10.3390/microorganisms9081562 (2021).Article 
PubMed 
PubMed Central 

Google Scholar 
Shen, J.-P., Chen, C. & Lewis, T. Long term repeated fire disturbance alters soil bacterial diversity but not the abundance in an Australian wet sclerophyll forest. Sci. Rep. 6, 19639. https://doi.org/10.1038/srep19639 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Williams, R. J., Hallgren, S. W. & Wilson, G. W. T. Frequency of prescribed burning in an upland oak forest determines soil and litter properties and alters the soil microbial community. For. Ecol. Manage. 265, 241–247. https://doi.org/10.1016/j.foreco.2011.10.032 (2012).Article 

Google Scholar 
Wang, Q., Zhong, M. & Wang, S. A meta-analysis on the response of microbial biomass, dissolved organic matter, respiration, and N mineralization in mineral soil to fire in forest ecosystems. For. Ecol. Manage. 271, 91–97. https://doi.org/10.1016/j.foreco.2012.02.006 (2012).Article 

Google Scholar 
Brockway, D. G., Gatewood, R. G. & Paris, R. B. Restoring fire as an ecological process in shortgrass prairie ecosystems: initial effects of prescribed burning during the dormant and growing seasons. J. Environ. Manage. 65, 135–152. https://doi.org/10.1006/jema.2002.0540 (2002).Article 
PubMed 

Google Scholar 
Deka, H. & Mishra, P. Effect of fuel burning on the microbial population of soil. Folia Microbiol. 29, 330–336 (1984).Article 

Google Scholar 
Weber, C., Lockhart, J., Charaska, E., Aho, K. & Lohse, K. Bacterial composition of soils in ponderosa pine and mixed conifer forests exposed to different wildfire burn severity. Soil Biol. Biochem. 69, 242–250. https://doi.org/10.1016/j.soilbio.2013.11.010 (2014).Article 

Google Scholar 
Choromanska, U. & DeLuca, T. H. Microbial activity and nitrogen mineralization in forest mineral soils following heating: evaluation of post-fire effects. Soil Biol. Biochem. 34, 263–271. https://doi.org/10.1016/S0038-0717(01)00180-8 (2002).Article 

Google Scholar 
Saccá, M. L., Barra Caracciolo, A., Di Lenola, M. & Grenni, P. in Soil Biological Communities and Ecosystem Resilience. (eds Lukac, M., Grenni, P. & Gamboni, M.) 9–24 (Springer International Publishing, 2017).Maquia, I. S. et al. Mining the microbiome of key species from African savanna woodlands: Potential for soil health improvement and plant growth promotion. Microorganisms 8(9), 1291 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Pressler, Y., Moore, J. C. & Cotrufo, M. F. Belowground community responses to fire: meta-analysis reveals contrasting responses of soil microorganisms and mesofauna. Oikos 128, 309–327. https://doi.org/10.1111/oik.05738 (2019).Article 

Google Scholar 
Pérez-Valera, E. et al. Fire modifies the phylogenetic structure of soil bacterial co-occurrence networks. Environ. Microbiol. https://doi.org/10.1111/1462-2920.13609 (2017).Article 
PubMed 

Google Scholar 
Baldrian, P. et al. Active and total microbial communities in forest soil are largely different and highly stratified during decomposition. ISME J. 6, 248–258. https://doi.org/10.1038/ismej.2011.95 (2012).Article 
PubMed 

Google Scholar 
Kobziar, L. N. et al. Pyroaerobiology: The aerosolization and transport of viable microbial life by wildland fire. Ecosphere 9, e02507. https://doi.org/10.1002/ecs2.2507 (2018).Article 

Google Scholar 
Carini, P. et al. Relic DNA is abundant in soil and obscures estimates of soil microbial diversity. Nat. Microbiol. 2, 16242. https://doi.org/10.1038/nmicrobiol.2016.242 (2016).Article 
PubMed 

Google Scholar 
Lennon, J. T., Muscarella, M. E., Placella, S. A. & Lehmkuhl, B. K. How, when, and where relic DNA affects microbial diversity. MBio 9, e00637-00618. https://doi.org/10.1128/mBio.00637-18 (2018).Article 

Google Scholar 
Leff, J. W. et al. Consistent responses of soil microbial communities to elevated nutrient inputs in grasslands across the globe. Proc. Natl. Acad. Sci. 112, 10967–10972 (2015).Article 
PubMed 
PubMed Central 

Google Scholar 
Dolan, K. L., Peña, J., Allison, S. D. & Martiny, J. B. Phylogenetic conservation of substrate use specialization in leaf litter bacteria. PLoS ONE 12, e0174472 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Woolet, J. & Whitman, T. Pyrogenic organic matter effects on soil bacterial community composition. Soil Biol. Biochem. 141, 107678 (2020).Article 

Google Scholar 
Buscardo, E. et al. Spatio-temporal dynamics of soil bacterial communities as a function of Amazon forest phenology. Sci. Rep. 8, 1–13 (2018).Article 

Google Scholar 
Tláskal, V., Zrůstová, P., Vrška, T. & Baldrian, P. Bacteria associated with decomposing dead wood in a natural temperate forest. FEMS Microbiol. Ecol. 93, fix157 (2017).Article 

Google Scholar 
Shade, A. & Handelsman, J. Beyond the Venn diagram: the hunt for a core microbiome. Environ. Microbiol. 14, 4–12. https://doi.org/10.1111/j.1462-2920.2011.02585.x (2012).Article 
PubMed 

Google Scholar  More

Study areaFigure 2 shows the location of the study area on a map of China generated by ArcGIS software. This study’s field experiments were carried out in the Shuanghe Town agricultural comprehensive water-saving demonstration area (40°42′ N; 107°24′ E), which is located in the middle reaches of the Hetao Irrigation Area of Inner Mongolia. The duration of the experimental process ranged from April in 2018 to October in 2020. The experimental area was characterized by a mid-temperate semi-arid continental climate. The average annual precipitation was determined to be 138 mm and the average evaporation was approximately 2332 mm. The majority of the rainfall was concentrated during summer and autumn seasons, and the accumulation of salt in the surface soil was considered to be serious in the spring and winter months. The average rainfall during maize growth period was 75.3 mm. The 0 to 40 cm soil layers in the experimental area were categorized as silty loam soil, with an average bulk density ranging from 1.42 to 1.53 g cm−3. A maize straw layer with a thickness of 5 cm was buried at a depth of 40 cm, and then the land was leveled. Also, in addition to autumn watering and spring irrigation procedures, water from the Yellow River was used three times for irrigation during the entire growth period of the maize crops. The adopted irrigation method belonged to border irrigation. Urea (46% N) were used as the fertilizer types.Figure 2The location of the study area.Full size imageField trials design and data collectionWe carried out experiment 1 from 2018 to 2019, and the data obtained were used for model training and to determine the hyper-parameters. The experimental design is shown in Table 1. The PNN model trained from the data obtained in experiment 1 predicted the optimal range of irrigation amount and nitrogen application rate (N rate) for each growth period of maize. In these ranges, the soil organic matter and total nitrogen could be kept above 20 g/kg and 1.6 g/kg, respectively, the soil salt content was less than 2 g/kg, and the pH value was between 6.5 and 7.5. In order to verify the accuracy and feasibility of the range of irrigation and nitrogen application simulated by PNN, the field experiment 2 was set in 2020 based on the range simulated by PNN and to evaluate the fitting degree between measured and simulated values of soil indicators under the same amount of irrigation and nitrogen application. The experimental design is shown in Table 2.Table 1 Experimental 1 design scheme.Full size tableTable 2 Experimental 2 design scheme.Full size tableThe experimental design were repeated for three times. The plot area of each treatment measuring 8 × 9 = 72 m2. The surrounding area was separated using 1.2 m buried polyethylene plastic film, and 30 cm was left at the top to prevent fertilizer and water from flowing into each other. The field management process was consistent with that used by the local farmers. The film width of maize was 1.1 m, with each film covering two rows. The plant spacing was approximately 45 cm, and the row spacing was 35 cm. In addition, the planting density of the maize was 60,000 plants/hm2.During the entire growth period of the maize crops, soil samples were collected from the 0 to 20 cm, 20 to 40 cm, 40 to 60 cm, 60 to 80 cm, and 80 to 100 cm soil layers using a soil drill and a three-point method was adopted. The soil samples were stored at 4 °C for the determination of total nitrogen, organic matter, total salt content, and pH values. The total nitrogen, organic matter, total salt content, and pH were determined using a KDN-AA double tube azotometer, MWD-2 microwave universal digestion device, TU1810PC ultraviolet–visible spectrophotometer, and a TU18950 double beam ultraviolet–visible spectrophotometer, respectively.Soil parameters measured include organic matter (SOM), total nitrogen (TN), Salt and pH. The data set includes pre-irrigation and post-irrigation reports from 2018 to 2020. Statistical parameters regarding the soil data are shown in Table 3.Table 3 Various meteorological variables and their descriptive statistics.Full size tableThe dataset obtained in Experiment 1 in 2018 to 2019 was 2490 rows in size, the 80/20 principle was used to data into training, and testing sets were required for ML modeling; 80% of data were employed for model training, while the remaining 20% were used for testing. Specifically, the data corresponding to the treatments with the nitrogen application rate (N rate) of 75 kg/hm2 (N3) in all the treatments (W1N3, W2N3, W3N3) were used as the test set, and the data of the other treatments were used as the training set. The training set was used to initiate ML parameter training. Subsequently, The test set was employed to assess the model. The dataset size in 2020 was 1080 rows, which was used to verify ML modeling.Figure 3 shows the changes of soil indexes over time for each treatment in the field test (take the 0–40 cm soil in the main distribution area of maize roots as an example). There are differences under the influence of different irrigation amounts. When irrigation is 90 mm, soil SOM is 13.25% and 7.00% higher than 60 mm and 120 mm, and soil TN is 4.59% and 6.50% higher than 60 mm and 120 mm, respectively. The soil Salt was 23.30% lower than 60 mm, and the pH was 4.16% and 4.36% lower than that of 60 mm and 120 mm, respectively. It can be seen that irrigation of 90 mm is more favorable for increasing soil SOM and TN contents and reducing soil salinity and alkalinity. Soil SOM and TN contents were the highest at n 75 kg/hm2, which were 4.38% and 8.34% higher than those at N = 93.3 kg/hm2, respectively. Soil Salt was the lowest at N = 60 kg/hm2, which was 3.02% lower than those at N = 75 kg/hm2, with a small gap with other levels. In conclusion, nitrogen application of 75 kg/hm2 was beneficial to increase soil organic matter and nitrogen content, and nitrogen application of 60 kg/hm2 was beneficial to controlling soil salt content.Figure 3Changes in soil organic matter, total nitrogen, salinity, and pH under different treatments over time (a case study of 2019).Full size imageMachine learning (ML) models used for irrigation and nitrogen application strategiesFive ML frames were used to estimate the irrigation and N rate. These models are preference Neural Network (PNN), Support Vector Regression (SVR), Linear Regression (LR), Logistic Regression (LOR), and traditional BP Neural Networks (BPNN). Among them, the prediction effects of linear, Poly, and rbf kernel functions are respectively tried in SVR framework. The torch framework was used to train and test machine learning models in Python.Development of preference neural networkModel frameworkThe preference neural network (PNN) which was proposed for the first time in this study was a typical deep learning model. PNN can be regarded as an approximate natural function in order to describe the complete dependence of the soil fertility indexes, including the effects of soil total nitrogen, organic matter, total salt content, and pH values on irrigation and nitrogen applications. More specifically, PNN has the ability to optimize the function by constructing the mapping y = f (x, θ) and learning parameter θ.First, the input end of PNN model was defined as matrix X ∈ ℝn×d (in which n is the sample size, n = 2490; and d is the dimension of each input vector, d = 6), where {xi} i=1, …, n ∈ X represents the vectorized set of total nitrogen, organic matter, salt content, and pH used for measuring the soil fertility, as well as the nitrogen application and irrigation durations (expressed by days after sowing). At the same time, the output end of the model was defined as the matrix Y ∈ ℝn×2, which represented the levels of the irrigation and nitrogen fertilizer applications. The goal of the proposed PNN model was to learn the fixed mapping Y′ = f (X; θ) ⇒Y through the given input matrix X, where θ is the well optimized learnable parameters which can be obtained via PNN training. Meanwhile, the predicted value Y′ will infinitely approach the measured value Y. The structure and the algorithm of this study’s PNN model is shown in Fig. 4 and Table. 4.Figure 4Schematic diagram for the PNN structural connections. In the figure, it can be seen that when each input vector passed through each layer of the PNN, it is first multiplied by the Hadamard product of the weight matrix and preference value matrix for the purpose of obtaining a weight matrix with preference properties. After the matrix was activated by the Relu Function, Batch Normalization Module Methods and the Dropout Module were used for random suspension and normalization processing, and the input of the next layer was obtained.Full size imageTable 4 Algorithm of Preference neural network.Full size tableLayer-by-layer affine transformationA good definition of the affine transformation of the information flow between layers is considered to be the key to neural network model training. Generally speaking, the learnable parameter θ of each layer of a model includes the weight parameter w and the preference parameter b. The hidden representation hl of the l-th layer in PNN is defined as follows:$${h}_{l}({h}_{l-1};{W}_{l},{b}_{l})={h}_{l-1}^{mathrm{T}}{W}_{l}+{b}_{l}$$
(1)

where Wl and bl represent the learnable weight and bias variables of the l layer, respectively, and hl-1 is the hidden representation of the upper layer. Therefore, when l = 1, then h0 = X.In the present study, using the hierarchical update rules, a given input data stream was allowed to pass through each hidden layer with intermediate operations, and then finally reached the output end.Preference structureThe correlation between different production behavior factors (e.g., irrigation levels) and different natural factors (e.g., soil organic matter) differs in agricultural production. However, the traditional fully connected neural network has the characteristic that nodes of one layer are fully connected with all nodes of subsequent layers, resulting in the neurons between production behavior factors and natural factors with very weak correlation still all being connected. Conversely, connections between neurons corresponding to factors with solid correlations are not strengthened.Therefore, in this study the preference value module was specially developed. By first calculating the correlation and significance between different production behavior factors (irrigation amount, N rate) and different soil fertility factors (organic matter, total nitrogen, total salt and pH), the preference value between the above two types of variables was calculated, and the preference matrix was constructed. Then the Hadamard product of the weight matrix and preference matrix was used to realize the artificial intervention and guidance to the neural network’s learning process.In order to reduce the adverse impact of non-normality of data on correlation analysis as much as possible, this study rank-based inverse normal (RIN) transformations (i.e., conversion to rank score) methods were used to normally process the data28. The RIN transformation function used here is as follows:$$f(x)={Phi }^{-1}left(frac{{x}_{r}-frac{1}{2}}{n}right)$$
(2)

where Φ–1 is the inverse normal cumulative distribution function, and n is the sample size.The normal cumulative distribution function is represented as follows: for discrete variables, the sum of probabilities of all values less than or equal to a, and its formula is as shown below:$${F}_{X}(a)=P(Xle a)$$
(3)
The RIN normalized conversion values meet the requirements of normal distribution, Pearson correlation analysis and t-test can be directly performed, and the formula used was as follows:$$r(X,Y)=frac{mathrm{Cov}(X,Y)}{sqrt{left(mathrm{Var}left[Xright]mathrm{Var}left[mathrm{Y}right]right)}}$$
(4)

where r (X, Y) is the Pearson Correlation Coefficient, Var [X] is the variance of X, and Var [Y] is the variance of Y, Cov (X, Y) is the covariance of X and Y, which represents the overall error of the two variables. The t-test is performed on the normalized data after rank-based inverse normal (RIN) transformation method, and the formula is as follows:$$t=sqrt{frac{n-2}{1-{r}^{2}}}$$
(5)

where n is the number of samples, and r represents the Pearson Correlation Coefficient. Preference value is the concentrated embodiment of correlation and significance between variables, and the calculation formula is as follows:$${PV}_{ij}=frac{r({X}_{i},{Y}_{j})}{{P}_{ij}+e}$$
(6)

where PVij represents the preference values between the variables Xi and Yj, Xi represents the ith production behavior factor (e.g., irrigation amount), and Yj represents the jth soil fertility factor (e.g., soil organic matter content), ({P}_{ij}) is obtained by looking up the table based on the t, and e is a constant, taking 0.001 in order to prevent the denominator of the formula from being 0.In order to make the preference values of the various indicators in the same order of magnitude more stable, the preference values were normalized:$${PV}_{normal}=pm frac{left|{PV}_{i}-{PV}_{avg}right|}{sqrt{frac{sum_{i=1}^{N}{({PV}_{i}-{PV}_{avg})}^{2}}{N-1}}}$$
(7)

where N represents the number of variables related to the experimental treatments, PVi -PVavg takes the absolute value, while the positive or negative values of the PVnormal were determined by the positive or negative values of the correlation r.The PNN integrated the preference matrixes into the neural network structures by identifying the Hadamard products of the learnable weights between the preference matrixes and the input and output data. By referring to Eq. (1) in the hierarchical affine transformation, the preference constraint of PNN could be expressed as follows:$${h}_{l}({h}_{l-1};{W}_{l},{b}_{l})={h}_{l-1}^{T}{W}_{l}odot P+{b}_{l}$$
(8)

where P is the preference matrix calculated by Eq. (8), and ⊙ represents the Hadamard product of the corresponding elements of the matrix. The structure of preference neural network and preference value are shown in Figs. 5 and 6.Figure 5Schematic diagram of the preference connection structures of the preference neural networks. The depth of the network detailed in the figure only illustrates the preference connection structure (for a better demonstration), and does not indicate the depth of the PNN used in the experiment.Full size imageFigure 6PVnormal between production behavior factors and natural factors. Since soil depth, days, irrigation amount and N rate were all artificially set variables, and there was no objective correlation in the data set. Therefore, the preference values among these variables were default e = 0.001.Full size imageHyper-parameters of PNNWe conducted experiments on the datasets with varying the hyper-parameters (such as the number of PNN layers and hidden layers, the number of nodes in each layer, learning rate, dropout rate and batch size) to understand that how the Hyper-parameters impact on the performance of PNN.We select the activation function and learning rate by referring to the neural network structure commonly used in similar fields (1 hidden layer and 64 hidden nodes)29,30. It is found that ReLU has better performance than other activation functions (sigmoid, tanh). The performance is best when the learning rate is around 0.005. It is generally believed that neural networks with more hidden layers are able, with the same number of resources, to address more complex problems31, but excessively increasing network depth will easily lead to overfitting32. Since there is no direct method to select the optimal number of hidden layers and nodes33, this study first calculated the structure of one hidden layer and 64 nodes in each layer, and found that the combined effect was poor (R2 of irrigation and nitrogen application were 0.3971 and 0.4124, respectively). Therefore, the trial-and-error method is adopted. The number of hidden layers starts from 1 and is incremented by 1 to test the maximum number of 10 hidden layers. The number of nodes in each layer were tested with a maximum number of 100 hidden neurons, starting with 5 and increasing by 5.We found that when the number of hidden layers of PNN exceeds 6, and the number of nodes in each layer exceeds 65, the performance will drop significantly. The reason behind this phenomenon could be the current dataset size is insufficient for larger scale of the PNN model. In the consideration of that the size of new dataset we can obtain very year is similar to the current dataset size, we believe that current hyper-paramter settings of PNN is in a reasonable condition.After that, the number of layers was fixed as 6, and the number of nodes in each layer were tested 10 times with 60 as the starting point and 1 as the increment, we found that when the number of nodes was 64, the improvement of the fit degree was no longer noticeable. On this basis, we changed different activation functions and learning rate again, and found that PNN still has the best performance when the activation function is ReLU and the learning rate is 0.005. Then, different batch sizes and dropout rates were tried. The two parameters had weaker effects on the performance than the other parameters, and the performance was optimal at 256 and 0.1, respectively.The hyper-parameters include:

1.

number of PNN layers;

2.

number of hidden layers;

3.

types of activation function;

4.

percentage of dropout;

5.

learning rate;

6.

loss function;

7.

optimizer;

8.

batch size;

9.

number of epochs;

10.

number of workers.

The ideal PNN structure for the study comprises these layers:

1.

number of PNN layers is 8;

2.

number of hidden layers is 6;

3.

Fully connected layers with 64 nodes and ReLU activation function

4.

dropout with 0.1.

5.

the learning rate is 0.005;

6.

loss function is Huber Loss Methods (HLM);

7.

optimizer: ADAM;

8.

epochs is 500;

9.

the batch size is 256;

10.

number of workers is 6.

Hyper-parameters of other modelsLR algorithms and LOR do not have hyper-parameters that need to be adjusted. A part of the hyper-parameters of the SVR model was determined by referring to Guan Xiaoyan’s research34, and a part of the hyper-parameters of the BPNN model was determined by referring to Gu Jian’s research27. RMLP takes the same hyperparameters as PNN. The hyperparameters of SVR and BPNN models are shown in Table 5.Table 5 Hyper-parameters of other model.Full size tableModel performance evaluationThe proposed PNN model was trained and validated using the field measured data from 2020 and the performance achievements of PNN were evaluated by the root mean square errors, mean square errors, and mean absolute errors as follows:$$RMSE=sqrt{frac{{sum }_{i=1}^{n}{({y}_{ipre}-{y}_{imea})}^{2}}{n}}$$
(9)
$${R}^{2}=1-frac{{sum }_{i=1}^{n}{({y}_{ipre}-{y}_{imea})}^{2}}{{sum }_{i=1}^{n}{({y}_{ipre}-{y}_{iavg})}^{2}}$$
(10)
$$MAE=frac{{sum }_{i=1}^{n}left|{y}_{ipre}-{y}_{iavg}right|}{n}$$
(11)
Model multidimensional fertility targetsThe soil fertility grade classification of soil organic matter, soil total nitrogen content and salt content in this study was based on the soil fertility grade classification results by the Agriculture and Animal Husbandry Bureau of Bayannur City, along with the local standard Technical Specifications for the Assessment and Rating Criteria of Cultivated Land Quality (DB 15/T 1086, 2016), as the shown in Tables 6 and 7.Table 6 Soil organic matter and Soil total nitrogen degrees.Full size tableTable 7 Grading of the salinization degrees.Full size tableIn the evaluation system of soil fertility referencing the Technical Specifications for Assessment and Rating Criteria of Cultivated Land Quality (DB 15/T 1086, 2016), the pH was divided into four grades according to the membership degrees of the land productivity evaluations, as detailed in Table 8.Table 8 pH grading degrees of the cultivated land.Full size tableBased on the classification standard of soil fertility obtained by the Bureau of Agriculture and Animal Husbandry of Bayannur City, when the farmland soil is at the high fertility level, the soil organic matter and total nitrogen content should be more than 20 g/kg and 1.6 g/kg, respectively. Soil salt content was less than 2 g/kg. Meanwhile, the pH value is kept between 6.5 and 7.5. More

Knowing the distribution of species is crucial for effective conservation action. However, accurate and high-resolution spatial data are only available for a limited number of species1,2. For mammals and birds, the most comprehensive and widely used global distribution dataset is the set of range maps compiled as part of the assessments for the International Union for Conservation of Nature (IUCN) Red List. These represent each species’ distributional limits and tend to minimize omission errors (i.e. false absences) at the expense of commission errors (i.e. false presences)3,4. Therefore, they often contain sizeable areas not regularly occupied by the species.Maps of the Area of Habitat (AOH; previously known as Extent of Suitable Habitat, ESH) complement range maps by indicating potential occupancy within the range, thereby reducing commission errors5. AOH is defined as ‘the habitat available to a species, that is, habitat within its range’5. These models are produced by subtracting areas unsuitable for the species within their range, using information on each species’ associations with habitat and elevation5,6,7,8. Comprehensive sets of AOH maps have been produced in the past for mammals6 and amphibians7, as well as subsets of birds8,9. The percentage of a species’ range covered by the AOH varies depending on the methodology used to associate species to their habitats, and their habitats to land-cover, the coarseness of the range map, the region in which the species is distributed, and the species’ habitat specialization and elevation limits5. For example, Rondinini et al.6 found that, when considering elevation and land cover features for terrestrial mammals, the AOH comprised, on average, 55% of the range. Ficetola et al.7 obtained a similar percentage when analyzing amphibians (55% for forest species, 42% for open habitat species and 61% for habitat generalists). Beresford et al.8 found that AOH covered a mean of 27.6% of the range maps of 157 threatened African bird species. In 2019, Brooks et al.5 proposed a formal definition and standardized methodology to produce AOH, limiting the inputs to habitat preferences, elevation limits, and geographical range.AOH production requires knowledge of which habitat types a species occurs in and their location within the range1. Information on habitat preference is documented for each species assessed in the IUCN Red List10, following the IUCN Habitats Classification Scheme11. However, the IUCN does not define habitat classes in a spatially explicit way, therefore, we used a recently published translation table that associates IUCN Habitat Classification Scheme classes with land cover classes12. Species’ elevation limits were also extracted from the IUCN Red List.We developed AOH maps for 5,481 terrestrial mammal species and 10,651 terrestrial bird species (Fig. 1). For 1,816 bird species defined by BirdLife International as migratory, we developed separate AOH maps, for the resident, breeding, and non-breeding ranges, according to the migratory distribution of the species (Fig. 2). The maps are presented in a regular latitude/longitude grid with an approximate 100 m resolution at the equator. On average, the AOH covers 66 ± 28% of the geographical range for mammals and 64 ± 27% for birds. We used the resulting AOH maps to produce four global species richness layers for: mammals, birds, globally threatened mammals and globally threatened birds13 (Fig. 3).Fig. 1Spatial distribution maps of Tangara abbas. Maps represent (a) the geographic range21, and (b) the Area of Habitat (AOH) of the species. The AOH was produced by subtracting unsuitable habitats from the geographical range. This species’ habitats are forest and terrestrial artificial habitats and has elevation range of 0 – 1600 m.Full size imageFig. 2Spatial distribution maps of Cardellina rubrifrons, divided into resident, breeding and non-breeding areas for this migratory species. Maps represent (a) the geographic range21, and (b) the Area of Habitat (AOH) of the species. The AOH was produced by subtracting unsuitable habitats from the ranges. This species is a forest species with elevation rangelimits of 1500 – 3100 m.Full size imageFig. 3Global species richness maps for (a) terrestrial mammals (considering 5,481species) and (b) terrestrial birds (considering 10,651 species). Calculated by overlaying all species’ AOH per class, resulting inon the number of species at each grid cell, latitude/longitude grid at a resolution of 1°/1008 or approximately 100 m at the equator (EPSG:4326) with the ellipsoid WGS 1984.Full size imageThe AOH maps presented in this paper are more useful for some purposes than global species distribution models, as they reduce and standardize commissions14. They are especially useful for not well-known and wide-range species. However, we note that for well-known species alternative sources may have more accurate distributions15. Moreover, AOHs are affected by the bias and errors of the underlying data, especially relevant errors associated with documentation of species’ habitats and elevations, and the translation of habitats into land cover classes, given that habitat is a complex multidimensional concept that is challenging to match to land-cover classes12, and that the current version of the IUCN Habitat Classification Scheme on IUCN’s website is described as a draft version11.The AOH maps have multiple conservation applications5,16,17, such as assessing species’ distributions and extinction risk, improving the accuracy of conservation planning, monitoring habitat loss and fragmentation, and guiding conservation actions. AOH has been proposed as an additional spatial metric to be documented in the Red List5, and is used for the identification of Key Biodiversity Areas18. More

Study sites and study designThe study was performed in the frame of a dog ecology research project, with details on the study locations published elsewhere15,42,43. For the current study, five study sites located in Indonesia and Guatemala were included. Site selection was carried out by each country’s research team, taking into consideration rural and urban settings, as well as differing expected number of dogs present at each location. The Indonesian study sites were semi-urban Habi and rural Pogon, in the Sikka regency, at the eastern area of Flores Island (Supplementary Fig. 6). In Guatemala, the study sites were Poptún (urban setting), Sabaneta and La Romana (both rural settings), located in the Guatemalan department of Péten, in the northern part of the country (Supplementary Fig. 7). Data were collected during May to June 2018 in Guatemala and from July to September 2018 in Indonesia.In each location, a 1 km2 area was predefined using Google Earth within which the study took place. The 1 km2 area was chosen because of the research goals of another part of the project, investigating the contact network of the dogs15. Within these areas, the teams visited all dog-owning households. In each household, the study was presented to an adult of the family, who was then asked if they owned a dog and if they were willing to participate in the study. After the dog owner’s oral or written consent was granted, a questionnaire was answered, and the dogs collared. The handling of the dogs was performed by a trained veterinarian or a trained veterinary paramedic of the team.The questionnaire data was collected through interviews with the dog owners. Multiple dogs per household could be included as multiple entries in the questionnaire. The detailed questionnaire contains information on the household location, dog demographics (age, sex, reproductive status) and management (dog’s purpose, origin, confinement, vaccination status, feeding and human-mediated transportation within and outside the pre-determined area).All dogs of a household fulfilling the inclusion criteria were equipped with a geo-referenced contact sensor (GCS) developed by Bonsai Systems (https://www.bonsai-systems.com), containing a GPS module and an Ultra-High-Frequency (UHF) sensor for contact data recording43,44. GCS devices report a 5-m maximum accuracy, a run-time of up to 10 years, can store up to 4 million data points and carry a lithium-polymer-battery (LiPo). For this study, only GPS data were analysed. The GCS were set to record each dog’s geographical position at one-minute intervals. Dogs remained collared for 3 to 5 days with the duration of the data collection being limited by the device’s battery capacity, as batteries were not re-charged or changed during the study. Throughout the time of recording, date, hour, GPS coordinates and signal quality (HDOP) raw data were collected by the GPS module and amassed into the workable databases.Exclusion criteria were dogs of less than four months of age (since they were not big enough to carry a collar), sick dogs and pregnant bitches (to avoid any risk of stress-induced miscarriages). Reasons for non-participation of eligible dogs included dog owner’s absence, dog’s absence, inability to catch the dog, and refusal of participation by the dog owner. In addition, dogs foreseen for slaughtering within the following four days were excluded in Indonesia to ensure data collection for at least four to five days. All dogs included in this study were constantly free roaming or at least part-time (day only, night only and for some hours a day). Human and/or animal ethical approval were obtained depending on the country-specific regulations. All the procedures were carried out in accordance with relevant guidelines. Ethical clearance was granted in Guatemala by the UVG’s International Animal Care and Use Committee [Protocol No. I-2018(3)] and the Community Development Councils of the two rural sites, which included Maya Q’eqchi’ communities45. In Indonesia, the study was approved by the Animal Ethics Commission of the Faculty of Veterinary Medicine, Nusa Cendana University (Protocol KEH/FKH/NPEH/2019/009). In addition, dogs that participated in the study were vaccinated against rabies and/or dewormed to acknowledge the owners for their participation in the study.Data cleaningData were stored in an application developed by Bonsai Systems compatible with Apple operating system (iOS iPhone Operating Systems), downloaded as individual csv file for each unit, and further analysed in R (version 3.6.1)46.The GPS data were cleaned based on three automatised criteria. First, the speed was calculated between any two consecutive GPS fixes, and fixes with speed of  > 20 km/h were excluded, given the implausibility of a dog running at such speed over a one-minute timespan47. It is noteworthy that car travel causes speeds over 20 km/h. However, as we were interested in analysing the dog’s behaviour outside of car transports, removing these fixes was in line with our objectives. Second, the Horizontal Dilution of Precision (HDOP), which is a measure of accuracy48 and automatically recorded by the devices for each GPS fix, was used to exclude fixes with low precision. According to Lewis et al.49, GPS fixes with HDOP higher than five were excluded, which deleted 1.3% of data in Habi, 2.2% in Pogon, 3.3% in Poptún, 1.8% in La Romana and 2.1% in Sabaneta. Third, the angles built by three consecutive fixes were calculated for each dog. When studying animals’ trajectories as their measure of movement, acute inner angles are often connected to error GPS fixes50. The fixes having the 2.5% smallest angles were excluded, to target those fixes with highest risks of being errors, while balancing against the loss of GPS fixes due to the cleaning process. With the exclusion of the smallest angles, 2.6% of data were deleted in Habi, 3% in Pogon, 2.9% in Poptún, 2.6% in La Romana and 2.7% in Sabaneta. After the automatised cleaning was concluded, 18 obvious error GPS fixes (unachievable or inexplicable locations by dogs) still prevailed in the Habi dataset and were manually removed.Habitat resource identification and calculation of terrain slopeTo analyse habitat selection of the collared FRDD, resources were delimited by a 100% Minimum Convex Polygon (MCP) including all cleaned GPS fixes per study site, using QGIS51 (Fig. 1).Figure 1GPS fixes plotted over a Google satellite imagery layer with its respective outlined computed Minimum Convex Polygon (MCP) delimitating the habitat available for the study population in: (a) Habi; (b) Pogon; (c) Poptún; (d) La Romana and (e) Sabaneta. Source QGIS (version 3.4 Madeira, http://qgis.org), map data: Google Satellite.Full size imageResources were defined by taking into consideration the following criteria: resources are (i) likely to impact upon movement patterns of dogs, (ii) identifiable by landscape satellite topography, and (iii) chosen considering information on relevant gathering places for FRDD observed by the field teams. Three resources were disclosed in all study sites: buildings, roads and vegetation coverage. All habitat relevant resources were manually identified within the available area (MCP) in QGIS using satellite imagery. All building-like structures were identified using vector polygons and summed under the layer “buildings”. Roads were identified and manually traced using vector lines in all sites, except in Poptún where the roads were automatically traced using an OpenStreetMap road layer of the area (https://www.openstreetmap.org/export). A buffer vector polygon was generated to encompass the full potential width of the roads, with a 5 m width in Habi and Poptún (semi-urban and urban site) and a 2 m width in Pogon, La Romana and Sabaneta (rural sites). In Habi, a “beach” layer was defined by generating a five-meter buffer from the shoreline in both directions using a vector polygon. The layer “sea” was defined as the vector polygon resulting from the difference between the MCP sea outer limit and the beach buffer polygon. Vegetation coverage was distinct between study sites with sparse vegetation and bushes present in all sites except Pogon, and dense forest-like vegetation present in La Romana and Pogon. These two types of vegetation were defined as “low” and “high vegetation”, respectively. In Habi and La Romana, “low” and “high vegetation”, respectively, were manually identified using vector polygons and summarised under the respective layers. Finally, open field in Habi, high vegetation in Pogon and low vegetation in Poptún, La Romana and Sabaneta were the last vector layers to be established since they represented the difference between all other polygon vector layers and the MCP total area. After all resource vector polygons had been created, an encompassing vector layer was generated by merging all resource polygon vectors for final resource classification (Fig. 2). As part of the resource classification in Habi, the airport terminal and runaway as well as waterways enclosed in the MCP area were identified but excluded from the analysis.Figure 2(a) Habi, (b) Pogon, (c) Poptún, (d) La Romana and (e) Sabaneta Habitat classification vector layers. The different habitat resources, identifiable by colour, were merged to create the comprehensive Habitat classification vector. In the Indonesian sites (a, b) and Guatemalan sites (c–e) buildings are coloured red, vegetation low in Habi, Poptún, La Romana and Sabaneta is coloured light green, vegetation high in Pogon and La Romana dark green, roads black, beach yellow, sea dark blue, airport grey, waterways light blue and open field light orange. The airport area (gray) and waterways (light blue) in Habi were not classified as separate habitat layers and were excluded from further analysis. Source QGIS (version 3.4 Madeira, http://qgis.org), map data: Google Satellite.Full size imageAfter the construction of the habitat resource layers, all GPS fixes were assigned to the respective resource they were located, using the QGIS join attributes by location algorithm. Fixes located exactly on the MCP border in Indonesia were not classified automatically and had to be manually classified to the respective resource.In non-flat topographies (all locations expect Habi) we tested the hypothesis of whether the steepness would influence the dogs’ movement patterns. The degrees of slope were calculated using a 30-m raster-cell resolution (STRM 1-Arc Second Global, downloaded from the United States Geological Survey (USGS) Earth Explorer, https://earthexplorer.usgs.gov/). The slope was assigned by the QGIS join attributes by location algorithm to each GPS fix.Statistical analysisTo quantify habitat selection in each study site, we compared resources used by the dogs with the resources available, according to Freitas et al.52. Adapting the methodology applied by O’Neill et al.18, the observed number of GPS fixes for each dog was used to generate an equivalent number of locations that were randomly distributed within the MCP area using the Random points in layer bound vector tool from QGIS. For example, if dog “D300” had 100 recorded GPS fixes, 100 random points were generated within the MCP of the respective study site and assigned to “D300”. Random points were then assigned to the respective resources and slope of that location, as previously done with the observed GPS fixes. Using this approach, the habitat resources used by each dog could be compared to the available resources in the respective study site, using a regression model.Observation independence is a fundamental presupposition of any regression model. However, the spatial nature of the point-referenced data permits perception of spatial dependence. In our dataset, spatial autocorrelation was proven for all study sites using the Moran’s I test. Therefore, we applied a spatial regression model, which takes into consideration spatial autocorrelation while exploring the effects of the study variables. A mixed effects logistic regression model accounting for spatial autocorrelation was created to quantify the effect of variables on used (i.e. observed GPS fix) versus available (i.e. randomly generated GPS fixes) resources, using the fitme function in the spaMM package in R53,54. The model’s binary outcome variable was defined as either observed (1) or random (0) GPS fix, i.e. the dog being present or absent from a position. The explanatory variable was the resource classification with “buildings”, “roads”, “low vegetation”, “beach”, “sea” and “open field” as levels in Habi; “buildings”, “roads” and “high vegetation” in Pogon; “buildings”, “roads”, “low vegetation” in Poptún and Sabaneta; and “buildings”, “roads”, and “high” and “low vegetation” in La Romana. Different habitat resources were used interchangeably as reference level. In all study sites except Habi, the slope was included as an additional explanatory variable. As observations were not evenly distributed in time, with less observations recorded towards the end of the study, a variable ”hour” was added as an additional continuous fixed effect.Each observed GPS fix was assigned to the hour of its record, with the earliest timestamp registered in each study site being assigned the hour zero. The randomly generated points were randomly assigned to an hour within the determined time continuum of the observed GPS fixes. As our focus was investigating habitat selection at a population-level, we assumed there was no within-dog autocorrelation (space/time) and each dog was independent and exhibited no group behaviour38. Still, to partially account for spatial autocorrelation of each dog’s household, the random effects included in models were defined as each dog’s household geographical location recorded during fieldwork by a GPS device. The restricted maximum likelihood (REML) through Laplace approximations, which can be applied to models with non-Gaussian random effects55, and the Matérn correlation function were used to fit the spatial models with the Matérn family dispersion parameter ν, indicator of strength of decay in the spatial effect, was set at 0.554. More

Pecl, G. et al. Biodiversity redistribution under climate change: impacts on ecosystems and human well-being. Science 355, eaai9214 (2017).Article 
PubMed 

Google Scholar 
Sunday, J. M., Bates, A. E. & Dulvy, N. K. Thermal tolerance and the global redistribution of animals. Nat. Clim. Change 2, 686–690 (2012).Article 

Google Scholar 
Pinsky, M. L., Eikeset, A. M., McCauley, D. J., Payne, J. L. & Sunday, J. M. Greater vulnerability to warming of marine versus terrestrial ectotherms. Nature 569, 108–111 (2019).Article 
PubMed 

Google Scholar 
Lenoir, J. et al. Species better track climate warming in the oceans than on land. Nat. Ecol. Evol. 4, 1044–1059 (2020).Article 
PubMed 

Google Scholar 
Poloczanska, E. S. et al. Global imprint of climate change on marine life. Nat. Clim. Change 3, 919–925 (2013).Article 

Google Scholar 
Sanford, E., Sones, J. L., García-Reyes, M., Goddard, J. H. & Largier, J. L. Widespread shifts in the coastal biota of northern California during the 2014–2016 marine heatwaves. Sci. Rep. 9, 4216 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Molinos, J. G., Burrows, M. & Poloczanska, E. Ocean currents modify the coupling between climate change and biogeographical shifts. Sci. Rep. 7, 1–9 (2017).
Google Scholar 
Sunday, J. M. et al. Species traits and climate velocity explain geographic range shifts in an ocean‐warming hotspot. Ecol. Lett. 18, 944–953 (2015).Article 
PubMed 

Google Scholar 
Figueira, W. F., Curley, B. & Booth, D. J. Can temperature-dependent predation rates regulate range expansion potential of tropical vagrant fishes? Mar. Biol. 166, 73 (2019).Article 

Google Scholar 
Champion, C. & Coleman, M. A. Seascape topography slows predicted range shifts in fish under climate change. Limnol. Oceanogr. Lett. 6, 143–153 (2021).Article 

Google Scholar 
Roberts, S. M., Boustany, A. M. & Halpin, P. N. Substrate-dependent fish have shifted less in distribution under climate change. Commun. Biol. 3, 1–7 (2020).Article 

Google Scholar 
Engelhard, G. H., Righton, D. A. & Pinnegar, J. K. Climate change and fishing: a century of shifting distribution in North Sea cod. Glob. Change Biol. 20, 2473–2483 (2014).Article 

Google Scholar 
Twiname, S. et al. A cross‐scale framework to support a mechanistic understanding and modelling of marine climate‐driven species redistribution, from individuals to communities. Ecography 43, 1764–1778 (2020).Article 

Google Scholar 
Bates, A. E. et al. Defining and observing stages of climate-mediated range shifts in marine systems. Glob. Environ. Change 26, 27–38 (2014).Article 

Google Scholar 
Fogarty, H. E., Burrows, M. T., Pecl, G. T., Robinson, L. M. & Poloczanska, E. S. Are fish outside their usual ranges early indicators of climate‐driven range shifts? Glob. Change Biol. 23, 2047–2057 (2017).Article 

Google Scholar 
Jiguet, F. & Barbet‐Massin, M. Climate change and rates of vagrancy of Siberian bird species to Europe. Ibis 155, 194–198 (2013).Article 

Google Scholar 
Burrows, M. T. et al. Geographical limits to species-range shifts are suggested by climate velocity. Nature 507, 492–495 (2014).Article 
PubMed 

Google Scholar 
Peck, M. A. et al. Projecting changes in the distribution and productivity of living marine resources: a critical review of the suite of modeling approaches used in the large European project VECTORS. Estuar., Coast. Shelf Sci. 201, 40–55 (2016).Article 

Google Scholar 
Brito-Morales, I. et al. Climate velocity can inform conservation in a warming world. Trends Ecol. Evol. 33, 441–457 (2018).Article 
PubMed 

Google Scholar 
Pinsky, M. L., Worm, B., Fogarty, M. J., Sarmiento, J. L. & Levin, S. A. Marine taxa track local climate velocities. Science 341, 1239–1242 (2013).Article 
PubMed 

Google Scholar 
Champion, C., Hobday, A. J., Zhang, X., Pecl, G. T. & Tracey, S. R. Changing windows of opportunity: past and future climate-driven shifts in temporal persistence of kingfish (Seriola lalandi) oceanographic habitat within south-eastern Australian bioregions. Mar. Freshw. Res. 70, 33–42 (2019).Article 

Google Scholar 
Pinsky, M. L., Selden, R. L. & Kitchel, Z. J. Climate-driven shifts in marine species ranges: scaling from organisms to communities. Annu. Rev. Mar. Sci. 12, 153–179 (2020).Article 

Google Scholar 
Lonhart, S. I., Jeppesen, R., Beas-Luna, R., Crooks, J. A. & Lorda, J. Shifts in the distribution and abundance of coastal marine species along the eastern Pacific Ocean during marine heatwaves from 2013 to 2018. Mar. Biodivers. Rec. 12, 1–15 (2019).Article 

Google Scholar 
Wernberg, T. et al. An extreme climatic event alters marine ecosystem structure in a global biodiversity hotspot. Nat. Clim. Change 3, 78–82 (2013).Article 

Google Scholar 
Lenanton, R., Dowling, C., Smith, K., Fairclough, D. & Jackson, G. Potential influence of a marine heatwave on range extensions of tropical fishes in the eastern Indian Ocean—Invaluable contributions from amateur observers. Regional Stud. Mar. Sci. 13, 19–31 (2017).Article 

Google Scholar 
Leriorato, J. C. & Nakamura, Y. Unpredictable extreme cold events: a threat to range-shifting tropical reef fishes in temperate waters. Mar. Biol. 166, 1–10 (2019).Article 

Google Scholar 
Hobday, A. J. & Pecl, G. T. Identification of global marine hotspots: sentinels for change and vanguards for adaptation action. Rev. Fish. Biol. Fish. 24, 415–425 (2014).Article 

Google Scholar 
Pecl, G. T. et al. Redmap Australia: challenges and successes with a large-scale citizen science-based approach to ecological monitoring and community engagement on climate change. Front. Mar. Sci. 6, 349 (2019).Article 

Google Scholar 
Jacox, M. G., Alexander, M. A., Bograd, S. J. & Scott, J. D. Thermal displacement by marine heatwaves. Nature 584, 82–86 (2020).Article 
PubMed 

Google Scholar 
Brown, C. J. et al. Ecological and methodological drivers of species’ distribution and phenology responses to climate change. Glob. Change Biol. 22, 1548–1560 (2016).Article 

Google Scholar 
Fuchs, H. L. et al. Wrong-way migrations of benthic species driven by ocean warming and larval transport. Nat. Clim. Change 10, 1052–1056 (2020).Article 

Google Scholar 
Rooney, N., McCann, K. S. & Moore, J. C. A landscape theory for food web architecture. Ecol. Lett. 11, 867–881 (2008).Article 
PubMed 

Google Scholar 
Feary, D. A. et al. Latitudinal shifts in coral reef fishes: why some species do and others do not shift. Fish. Fish. 15, 593–615 (2014).Article 

Google Scholar 
Beissinger, S. R. & Riddell, E. A. Why are species’ traits weak predictors of range shifts? Ann. Rev. Ecol. Evol. Syst. 52, 47–66 (2021).Pearce, A. F. & Feng, M. The rise and fall of the “marine heat wave” off Western Australia during the summer of 2010/2011. J. Mar. Syst. 111, 139–156 (2013).Article 

Google Scholar 
Oliver, E. C. et al. The unprecedented 2015/16 Tasman Sea marine heatwave. Nat. Commun. 8, 16101 (2017).Article 
PubMed 
PubMed Central 

Google Scholar 
Gervais, C. R., Champion, C. & Pecl, G. T. Species on the move around the Australian coastline: a continental scale review of climate‐driven species redistribution in marine systems. Glob. Change Biol. 27, 3200–3217 (2021).Article 

Google Scholar 
Nursey-Bray, M., Palmer, R. & Pecl, G. Spot, log, map: assessing a marine virtual citizen science program against Reed’s best practice for stakeholder participation in environmental management. Ocean Coast. Manag. 151, 1–9 (2018).Article 

Google Scholar 
Pecl, G. T. et al. Ocean warming hotspots provide early warning laboratories for climate change impacts. Rev. Fish. Biol. Fish. 24, 409–413 (2014).Article 

Google Scholar 
Stuart-Smith, J. et al. Southernmost records of two Seriola species in an Australian ocean-warming hotspot. Mar. Biodivers. 48, 1579–1582 (2018).Article 

Google Scholar 
Provoost, P. & Bosch, S. robis: R Client to access data from the OBIS API. Ocean Biogeographic Information System. Intergovernmental Oceanographic Commission of UNESCO. R package version 2.1.8, https://cran.r-project.org/package=robis (2019).Froese, R. & Pauly, D. (eds). FishBase. World Wide Web electronic publication. www.fishbase.org. (2022). Accessed 14 July 2019.ABRS. Australian Faunal Directory. Australian Biological Resources Study, Canberra. https://biodiversity.org.au/afd/home. (2020). Accessed 15 July 2019.Robinson, L. M. et al. Rapid assessment of an ocean warming hotspot reveals “high” confidence in potential species’ range extensions. Glob. Environ. Change 31, 28–37 (2015).Article 

Google Scholar 
Hijmans, R. J. raster: geographic data analysis and modeling. R package version 3.4-5. https://CRAN.R-project.org/package=raster (2020).van Etten, J. R package gdistance: distances and routes on geographical grids. J. Stat. Softw. 76, 1–21 (2017).
Google Scholar 
Hobday, A. J. et al. A hierarchical approach to defining marine heatwaves. Prog. Oceanogr. 141, 227–238 (2016).Article 

Google Scholar 
Molinos, J. G., Schoeman, D. S., Brown, C. J. & Burrows, M. T. VoCC: an r package for calculating the velocity of climate change and related climatic metrics. Methods Ecol. Evol. 10, 2195–2202 (2019).Article 

Google Scholar 
Schlegel, R. W. & Smit, A. J. heatwaveR: a central algorithm for the detection of heatwaves and cold-spells. J. Open Source Softw. 3, 821 (2018).Article 

Google Scholar 
Venables, W. N. & Ripley, B. D. Modern applied statistics with S. 4th edn, (Springer, 2002).Lüdecke, D. sjPlot: data visualization for statistics in social science. R package version 2.8.6. https://CRAN.R-project.org/package=sjPlot (2020). More

Rogers, A. et al. Antarctic futures: An assessment of climate-driven changes in ecosystem structure, function, and service provisioning in the southern ocean. Annual Review of Marine Science 12, 87–120, https://doi.org/10.1146/annurev-marine-010419-011028 (2020).Article 
PubMed 

Google Scholar 
Tin, T. et al. Impacts of local human activities on the antarctic environment. Antarctic Science 21, 3–33, https://doi.org/10.1017/S0954102009001722 (2009).Article 

Google Scholar 
Pineda-Metz, S. E. A., Gerdes, D. & Richter, C. Benthic fauna declined on a whitening antarctic continental shelf. Nature Communications 11, 2226, https://doi.org/10.1038/s41467-020-16093-z (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Convey, P. Antarctic terrestrial biodiversity in a changing world. Polar Biology 34, 1629, https://doi.org/10.1007/s00300-011-1068-0 (2011).Article 

Google Scholar 
Kang, Y. H. et al. Composition and structure of the marine benthic community in terra nova bay, antarctica: Responses of the benthic assemblage to disturbances. PLOS ONE 14, 1–16, https://doi.org/10.1371/journal.pone.0225551 (2019).Article 

Google Scholar 
Piazza, P. et al. Underwater photogrammetry in antarctica: long-term observations in benthic ecosystems and legacy data rescue. Polar Biology 42, 1061–1079, https://doi.org/10.1007/s00300-019-02480-w (2019).Article 

Google Scholar 
SOOS. Southern Ocean Observing System – Report on the 2017 Ross Sea Working Group Meeting. http://www.soos.aq. [Online; accessed 2022/11/15] (2017).SCAR. Scientific Committee on Antarctic Research. https://www.scar.org. [Online; accessed 2022/11/15] (2021).ANTOS. Antarctic near-shore and terrestrial observing system. https://www.scar.org/science/antos/home. [Online; accessed 2022/11/15] (2015).Dayton, P. K. et al. Benthic responses to an antarctic regime shift: food particle size and recruitment biology. Ecological Applications 29, e01823, https://doi.org/10.1002/eap.1823 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Watters, G. M., Hinke, J. T. & Reiss, C. S. Long-term observations from antarctica demonstrate that mismatched scales of fisheries management and predator-prey interaction lead to erroneous conclusions about precaution. Scientific Reports 10, 2314, https://doi.org/10.1038/s41598-020-59223-9 (2020).Article 
PubMed 
PubMed Central 

Google Scholar 
Bolinesi, F. et al. Spatial-related community structure and dynamics in phytoplankton of the ross sea, antarctica. Frontiers in Marine Science 7, https://doi.org/10.3389/fmars.2020.574963 (2020).Stenni, B. et al. Three-year monitoring of stable isotopes of precipitation at concordia station, east antarctica. The Cryosphere 10, 2415–2428, https://doi.org/10.5194/tc-10-2415-2016 (2016).Article 

Google Scholar 
Ramesh, K. & Soni, V. Perspectives of antarctic weather monitoring and research efforts. Polar Science 18, 183–188, https://doi.org/10.1016/j.polar.2018.04.005 (2018). Recent Advances in Climate Science of Polar Region (to commemorate the contributions of Late Dr. S.Z. Qasim, a pioneering doyen of the Indian Polar programme).Article 

Google Scholar 
Shepherd, A. et al. Mass balance of the antarctic ice sheet from 1992 to 2017. Nature 558, 219–222, https://doi.org/10.1038/s41586-018-0179-y (2018).Article 

Google Scholar 
Budge, J. S. & Long, D. G. A comprehensive database for antarctic iceberg tracking using scatterometer data. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 11, 434–442, https://doi.org/10.1109/JSTARS.2017.2784186 (2018).Article 

Google Scholar 
Rignot, E. et al. Four decades of antarctic ice sheet mass balance from 1979–2017. Proceedings of the National Academy of Sciences of the United States of America 116, 1095–1103, https://doi.org/10.1073/pnas.1812883116 (2019).Article 
PubMed 
PubMed Central 

Google Scholar 
Barbat, M. M., Rackow, T., Wesche, C., Hellmer, H. H. & Mata, M. M. Automated iceberg tracking with a machine learning approach applied to sar imagery: A weddell sea case study. ISPRS Journal of Photogrammetry and Remote Sensing 172, 189–206, https://doi.org/10.1016/j.isprsjprs.2020.12.006 (2021).Article 

Google Scholar 
Aguzzi, J. et al. New high-tech flexible networks for the monitoring of deep-sea ecosystems. Environmental Science & Technology 53, 6616–6631, https://doi.org/10.1021/acs.est.9b00409 (2019).Article 

Google Scholar 
Piazza, P., Gattone, S., Guzzi, A. & Schiaparelli, S. Towards a robust baseline for long-term monitoring of antarctic coastal benthos. Hydrobiologia 847, 1753–1771, https://doi.org/10.1007/s10750-020-04177-2 (2020).Article 

Google Scholar 
Rountree, R. et al. Towards an optimal design for ecosystem-level ocean observatories. Oceanography and Marine Biology 58, 79–105, https://doi.org/10.1201/9780429351495-2 (2020).Article 

Google Scholar 
Katsanevakis, S. et al. Monitoring marine populations and communities: Methods dealing with imperfect detectability. Aquatic Biology 16, 31–52, https://doi.org/10.3354/ab00426 (2012).Article 

Google Scholar 
Zampoukas, N. et al. Technical guidance on monitoring for the marine strategy framework directive. Tech. Rep., European Commission, Report EUR 26499 (2014).Bicknell, A. W., Godley, B. J., Sheehan, E. V., Votier, S. C. & Witt, M. J. Camera technology for monitoring marine biodiversity and human impact. Frontiers in Ecology and the Environment 14, 424–432, https://doi.org/10.1002/fee.1322 (2016).Article 

Google Scholar 
European Marine Board. Working Group on Big Data in Marine Science. https://www.marineboard.eu/publications/big-data-marine-science. [Online; accessed 2022/11/15] (2020).Zurowietz, M. & Nattkemper, T. W. Current trends and future directions of large scale image and video annotation: Observations from four years of biigle 2.0. Frontiers in Marine Science 8, https://doi.org/10.3389/fmars.2021.760036 (2021).Kim, S. L., Thurber, A., Hammerstrom, K. & Conlan, K. Seastar response to organic enrichment in an oligotrophic polar habitat. Journal of Experimental Marine Biology and Ecology 346, 66–75, https://doi.org/10.1016/j.jembe.2007.03.004 (2007).Article 

Google Scholar 
Peirano, A., Bordone, A., Marini, S., Piazza, P. & Schiaparelli, S. A simple time-lapse apparatus for monitoring macrozoobenthos activity in antarctica. Antarctic Science 28, 473–474, https://doi.org/10.1017/S0954102016000377 (2016).Article 

Google Scholar 
Peirano, A., Marini, S., Bordone, A. & Schiaparelli, S. ICE-LAPSE: Analysis of antarctic benthos dynamics by using non-destructive monitoring devices and permanent stations, pnra 2013/az1.16, funded by the italian national antarctic program (2015-2016).Marini, S. et al. Long-term automated visual monitoring of antarctic benthic fauna. Methods in Ecology and Evolution 13, 1746–1764, https://doi.org/10.1111/2041-210X.13898 (2022).Article 

Google Scholar 
Marini, S. et al. EP2863257 (A1) – Underwater images acquisition and processing system. https://data.epo.org/gpi/EP2863257B1. [Online; accessed 2022/11/15] (2013).Corgnati, L. et al. Looking inside the ocean: Toward an autonomous imaging system for monitoring gelatinous zooplankton. Sensors 16, 2124, https://doi.org/10.3390/s16122124 (2016).Article 
PubMed 
PubMed Central 

Google Scholar 
Marini, S. et al. Automated estimate of fish abundance through the autonomous imaging device guard1. Measurement 126, 72–75, https://doi.org/10.1016/j.measurement.2018.05.035 (2018).Article 

Google Scholar 
Pensieri, S. et al. Environmental acoustic noise observations in tethys bay (terra nova bay, ross sea, antarctica). In 2014 Oceans – St. John’s, 1–6, https://doi.org/10.1109/OCEANS.2014.7003196 (2014).Jung, J. et al. Multibeam bathymetry and distribution of clay minerals on surface sediments of a small bay in terra nova bay, antarctica. Minerals 11, https://doi.org/10.3390/min11010072 (2021).Balog, I. et al. Estimation of direct normal irradiance at antarctica for concentrated solar technology. Applied System Innovation 2, https://doi.org/10.3390/asi2030021 (2019).Caputi, S. S. et al. Seasonal food web dynamics in the antarctic benthos of tethys bay (ross sea): Implications for biodiversity persistence under different seasonal sea-ice coverage. Frontiers in Marine Science 7, 1046, https://doi.org/10.3389/fmars.2020.594454 (2020).Article 

Google Scholar 
van Leeuwe, M. A. et al. Annual patterns in phytoplankton phenology in antarctic coastal waters explained by environmental drivers. Limnology and Oceanography 65, 1651–1668, https://doi.org/10.1002/lno.11477 (2020).Article 

Google Scholar 
OEngineering. OEngineering s.r.l. – GUARD-1, Underwater Autonomous Smart Camera. https://www.oengineering.eu//GUARD-1/. [Online; accessed 2022/11/15] (2021).Magic Lantern. https://magiclantern.fm. [Online; accessed 2022/11/15] (2021).Marini, S. et al. Guard1: An autonomous system for gelatinous zooplankton image-based recognition. In OCEANS 2015 – Genova, 1–7, https://doi.org/10.1109/OCEANS-Genova.2015.7271704 (2015).CR2. The Canon RAW (CRW) File Format. https://exiftool.org/canon_raw.html. [Online; accessed 2022/11/15] (2022).Marini, S. et al. ICE-LAPSE image dataset. Zenodo https://doi.org/10.5281/zenodo.6418163 (2022).LabelImg. A graphical image annotation tool. https://github.com/tzutalin/labelImg. [Online; accessed 2022/11/15] (2021).Schoening, T. et al. Making marine image data fair. Scientific Data 9, 414, https://doi.org/10.1038/s41597-022-01491-3 (2022).Article 
PubMed 
PubMed Central 

Google Scholar 
Cattaneo-Vietti, R., Chiantore, M., Schiaparelli, S. & Albertelli, G. Shallow- and deep-water mollusc distribution at terra nova bay (ross sea, antarctica). Polar Biology 23, 173–182, https://doi.org/10.1007/s003000050024 (2000).Article 

Google Scholar 
Cattaneo-Vietti, R. et al. Spatial and Vertical Distribution of Benthic Littoral Communities in Terra Nova Bay, 503–514 (Springer Berlin Heidelberg, Berlin, Heidelberg, 2000).Cummings, V. J. et al. Linking ross sea coastal benthic communities to environmental conditions: Documenting baselines in a spatially variable and changing world. Frontiers in Marine Science 5, 232, https://doi.org/10.3389/fmars.2018.00232 (2018).Article 

Google Scholar 
Redmon, J., Divvala, S., Girshick, R. & Farhadi, A. You only look once: Unified, real-time object detection. In 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 779–788, https://doi.org/10.1109/CVPR.2016.91 (2016).YOLO V5. https://github.com/ultralytics/yolov5. [Online; accessed 2022/11/15] (2022). More

United Nations. World Urbanization Prospects: The 2018 revision. (Department of Economic and Social Affairs, Population Division, United Nations, 2018).Grimm, N. B. et al. Global change and the ecology of cities. Science 319, 756–760 (2008).Article 

Google Scholar 
McPhearson, T. et al. Advancing urban ecology toward a science of cities. Bioscience 66, 198–212 (2016).Article 

Google Scholar 
Dodman, D. et al. Cities, settlements and key infrastructure. In Climate Change 2022: Impacts, Adaptation, and Vulnerability. Contribution of Working Group II to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change (eds. Pörtner, H.-O. et al.) 997–1040 (Cambridge University Press, Cambridge, UK and New York, NY, USA, 2022).Díaz, S. et al. Assessing nature’s contributions to people: Recognizing culture, and diverse sources of knowledge, can improve assessments. Science 359, 270–272 (2018).Article 

Google Scholar 
Grabowski, Z. J., McPhearson, T., Matsler, A. M., Groffman, P. & Pickett, S. T. A. What is green infrastructure? A study of definitions in US city planning. Front. Ecol. Environ. 20, 152–160 (2022).Article 

Google Scholar 
Childers, D. L. et al. Urban ecological infrastructure: An inclusive concept for the non-built urban environment. Elementa 7, 1–14 (2019).
Google Scholar 
Gómez-Baggethun, E. et al. Urban ecosystem services. In Urbanization, Biodiversity and Ecosystem Services: Challenges and Opportunities (eds. Elmqvist, T. et al.) 175–251 (Springer, Netherlands, 2013).Díaz, S. & Cabido, M. Vive la différence: Plant functional diversity matters to ecosystem processes. Trends Ecol. Evol. 16, 646–655 (2001).Article 

Google Scholar 
Burkhard, B. & Maes, J. Mapping Ecosystem Services (Pensoft Publishers, Sofia, 2017).Eviner, V. T. & Chapin, F. S. Functional Matrix: A conceptual framework for predicting multiple plant effects on ecosystem processes. Annu. Rev. Ecol. Evol. Syst. 34, 455–485 (2003).Article 

Google Scholar 
Lavorel, S., McIntyre, S., Landsberg, J. & Forbes, T. D. A. Plant functional classifications: From general groups to specific groups based on response to disturbance. Trends Ecol. Evol. 12, 474–478 (1997).Article 

Google Scholar 
Cornelissen, J. H. C. et al. A handbook of protocols for standardised and easy measurement of plant functional traits worldwide. Aust. J. Bot. 51, 335–380 (2003).Article 

Google Scholar 
Suding, K. N. et al. Scaling environmental change through the community-level: A trait-based response-and-effect framework for plants. Glob. Chang. Biol. 14, 1125–1140 (2008).Article 

Google Scholar 
Lavorel, S. & Garnier, E. Predicting changes in community composition and ecosystem functioning from plant traits: Revisiting the Holy Grail. Funct. Ecol. 16, 545–556 (2002).Article 

Google Scholar 
Hevia, V. et al. Trait-based approaches to analyze links between the drivers of change and ecosystem services: Synthesizing existing evidence and future challenges. Ecol. Evol. 7, 831–844 (2017).Article 

Google Scholar 
Cadotte, M. W., Carscadden, K. & Mirotchnick, N. Beyond species: Functional diversity and the maintenance of ecological processes and services. J. Appl. Ecol. 48, 1079–1087 (2011).Article 

Google Scholar 
Lavorel, S. Plant functional effects on ecosystem services. J. Ecol. 101, 4–8 (2013).Article 

Google Scholar 
Andersson, E. et al. What are the traits of a social-ecological system: Towards a framework in support of urban sustainability. npj Urban Sustain. 1, 14 (2021).Article 

Google Scholar 
Pickett, S. T. A. et al. Urban ecological systems: Linking terrestrial ecological, physical, and socioeconomic components of metropolitan areas. Annu. Rev. Ecol. Syst. 32, 127–157 (2001).Article 

Google Scholar 
McPhearson, T., Haase, D., Kabisch, N. & Gren, Å. Advancing understanding of the complex nature of urban systems. Ecol. Indic. 70, 566–573 (2016).Article 

Google Scholar 
Zhou, W., Pickett, S. T. A. & McPhearson, T. Conceptual frameworks facilitate integration for transdisciplinary urban science. npj Urban Sustain. 1, 1 (2021).Article 

Google Scholar 
Andersson, E. et al. Scale and context dependence of ecosystem service providing units. Ecosyst. Serv. 12, 157–164 (2015).Article 

Google Scholar 
Pinho, P. et al. Research agenda on biodiversity and ecosystem functions and services in European cities. Basic Appl. Ecol. 53, 124–133 (2021).Article 

Google Scholar 
Bullock, J. M. et al. Human-mediated dispersal and the rewiring of spatial networks. Trends Ecol. Evol. 33, 958–970 (2018).Article 

Google Scholar 
Avolio, M. L., Swan, C., Pataki, D. E. & Jenerette, G. D. Incorporating human behaviors into theories of urban community assembly and species coexistence. Oikos 130, 1849–1864 (2021).Article 

Google Scholar 
Aronson, M. F. J. et al. Hierarchical filters determine community assembly of urban species pools. Ecology 97, 2952–2963 (2016).Article 

Google Scholar 
Woodward, F. I. & Diament, A. D. Functional approaches to predicting the ecological effects of global change. Funct. Ecol. 5, 212 (1991).Article 

Google Scholar 
Diaz, S., Cabido, M. & Casanoves, F. Plant functional traits and environmental filters at a regional scale. J. Veg. Sci. 9, 113–122 (1998).Article 

Google Scholar 
Boet, O., Arnan, X. & Retana, J. The role of environmental vs. biotic filtering in the structure of European ant communities: A matter of trait type and spatial scale. PLoS ONE 15, e0228625 (2020).Article 

Google Scholar 
Grimm, N. B., Grove, J. M., Pickett, S. T. A. & Redman, C. L. Integrated approaches to long-term studies of urban ecological systems. Bioscience 50, 571–584 (2000).Article 

Google Scholar 
Vandewalle, M. et al. Functional traits as indicators of biodiversity response to land use changes across ecosystems and organisms. Biodivers. Conserv. 19, 2921–2947 (2010).Article 

Google Scholar 
Williams, N. S. G. et al. A conceptual framework for predicting the effects of urban environments on floras. J. Ecol. 97, 4–9 (2009).Article 

Google Scholar 
Cavender-Bares, J. et al. Horticultural availability and homeowner preferences drive plant diversity and composition in urban yards. Ecol. Appl. 30, 1–16 (2020).Article 

Google Scholar 
Pearse, W. D. et al. Homogenization of plant diversity, composition, and structure in North American urban yards. Ecosphere 9, e02105 (2018).Article 

Google Scholar 
Cubino, J. P. et al. Drivers of plant species richness and phylogenetic composition in urban yards at the continental scale. Landsc. Ecol. 34, 63–77 (2019).Article 

Google Scholar 
Oke, T. R. The energetic basis of the urban heat island. Q. J. R. Meteorol. Soc. 108, 1–24 (1982).
Google Scholar 
Sukopp, H. Human-caused impact on preserved vegetation. Landsc. Urban Plan. 68, 347–355 (2004).Article 

Google Scholar 
Díaz, S. et al. Incorporating plant functional diversity effects in ecosystem service assessments. Proc. Natl. Acad. Sci. USA 104, 20684–20689 (2007).Article 

Google Scholar 
Williams, N. S. G., Hahs, A. K. & Vesk, P. A. Urbanisation, plant traits and the composition of urban floras. Perspect. Plant Ecol. Evol. Syst. 17, 78–86 (2015).Article 

Google Scholar 
Teskey, R. et al. Responses of tree species to heat waves and extreme heat events. Plant Cell Environ. 38, 1699–1712 (2015).Article 

Google Scholar 
Jochner, S. & Menzel, A. Urban phenological studies—past, present, future. Environ. Pollut. 203, 250–261 (2015).Article 

Google Scholar 
Cleland, E. E., Chuine, I., Menzel, A., Mooney, H. A. & Schwartz, M. D. Shifting plant phenology in response to global change. Trends Ecol. Evol. 22, 357–365 (2007).Article 

Google Scholar 
de Bello, F. et al. Towards an assessment of multiple ecosystem processes and services via functional traits. Biodivers. Conserv. 19, 2873–2893 (2010).Article 

Google Scholar 
Santangelo, J. S. et al. Global urban environmental change drives adaptation in white clover. Science 375, 1275–1281 (2022).Article 

Google Scholar 
Blois, J. L., Zarnetske, P. L., Fitzpatrick, M. C. & Finnegan, S. Climate change and the past, present, and future of biotic interactions. Science 341, 499–504 (2013).Article 

Google Scholar 
Martin, C. A., Warren, P. S. & Kinzig, A. P. Neighborhood socioeconomic status is a useful predictor of perennial landscape vegetation in residential neighborhoods and embedded small parks of Phoenix, AZ. Landsc. Urban Plan. 69, 355–368 (2004).Article 

Google Scholar 
Kinzig, A. P., Warren, P., Martin, C., Hope, D. & Katti, M. The effects of human socioeconomic status and cultural characteristics on urban patterns of biodiversity. Ecol. Soc. 10, 23 (2005).Article 

Google Scholar 
Stephenson, J. The cultural values model: An integrated approach to values in landscapes. Landsc. Urban Plan. 84, 127–139 (2008).Article 

Google Scholar 
Andersson, E., Barthel, S. & Ahrné, K. Measuring social–ecological dynamics behind the generation of ecosystem services. Ecol. Appl. 17, 1267–1278 (2007).Article 

Google Scholar 
Fraser, E. D. G. & Kenney, W. A. Cultural background and landscape history as factors affecting perceptions of the urban forest. J. Arboric. 26, 106–113 (2000).
Google Scholar 
Hope, D. et al. Socioeconomics drive urban plant diversity. Proc. Natl. Acad. Sci. USA 100, 8788–8792 (2003).Article 

Google Scholar 
Avolio, M. L. et al. Understanding preferences for tree attributes: The relative effects of socio-economic and local environmental factors. Urban Ecosyst. 18, 73–86 (2015).Article 

Google Scholar 
Körmöndi, B., Tempfli, J., Kocsis, J. B., Adams, J. & Szkordilisz, F. E. The secret ingredient—The role of governance in green infrastructure development: Through the examples of European cities. IOP Conf. Ser. Earth Environ. Sci. 323, (2019).Conway, T. M. & Vander Vecht, J. Growing a diverse urban forest: species selection decisions by practitioners planting and supplying trees. Landsc. Urban Plan. 138, 1–10 (2015).Article 

Google Scholar 
Lack, W. H. The Book of Palms (Taschen-Bibliotheca Universalis, 2015).Grilo, F. et al. Using green to cool the grey: Modelling the cooling effect of green spaces with a high spatial resolution. Sci. Total Environ. 724, 138182 (2020).Article 

Google Scholar 
Prasifka, J. R. et al. Using nectar-related traits to enhance crop–pollinator interactions. Front. Plant Sci. 9, 1–8 (2018).Article 

Google Scholar 
Veerkamp, C. J. et al. A review of studies assessing ecosystem services provided by urban green and blue infrastructure. Ecosyst. Serv. 52, 101367 (2021).Article 

Google Scholar 
Theodorou, P. et al. Urban areas as hotspots for bees and pollination but not a panacea for all insects. Nat. Commun. 11, 576 (2020).Article 

Google Scholar 
Farmer, J. Trees in Paradise: A California History (WW Norton & Company, 2013).Goodness, J., Andersson, E., Anderson, P. M. L. & Elmqvist, T. Exploring the links between functional traits and cultural ecosystem services to enhance urban ecosystem management. Ecol. Indic. 70, 597–605 (2016).Article 

Google Scholar 
Masterson, V. A. et al. The contribution of sense of place to social-ecological systems research: A review and research agenda. Ecol. Soc. 22, 49 (2017).Article 

Google Scholar 
Masterson, V. A., Enqvist, J. P., Stedman, R. C. & Tengö, M. Sense of place in social-ecological systems: From theory to empirics. Sustain. Sci. 14, 555–564 (2019).Article 

Google Scholar 
Mukherjee, A. & Agrawal, M. Use of GLM approach to assess the responses of tropical trees to urban air pollution in relation to leaf functional traits and tree characteristics. Ecotoxicol. Environ. Saf. 152, 42–54 (2018).Article 

Google Scholar 
Singh, S. K., Rao, D. N., Agrawal, M., Pandey, J. & Naryan, D. Air pollution tolerance index of plants. J. Environ. Manage. 32, 45–55 (1991).Article 

Google Scholar 
Mukherjee, A. & Agrawal, M. Pollution response score of tree species in relation to ambient air quality in an urban area. Bull. Environ. Contam. Toxicol. 96, 197–202 (2016).Article 

Google Scholar 
Barwise, Y. & Kumar, P. Designing vegetation barriers for urban air pollution abatement: A practical review for appropriate plant species selection. npj Clim. Atmos. Sci. 3, 12 (2020).Article 

Google Scholar 
Grote, R. et al. Functional traits of urban trees: Air pollution mitigation potential. Front. Ecol. Environ. 14, 543–550 (2016).Article 

Google Scholar 
Tomson, M. et al. Green infrastructure for air quality improvement in street canyons. Environ. Int. 146, 106288 (2021).Article 

Google Scholar  More

Study area descriptionYangtze River Basin is situated in central China (Fig. 1). Its geographical coordinates are between 30° 48′ 30″–31° 02′ 30″ N and 112° 48′ 45″–113° 03′ 45″ E. Taizishan is located in the transition zone between the north and south of China, with an altitude of 403–467.4 m. It belongs to the subtropical monsoon humid climate zone and has obvious karst landforms. The farm area is 7576 hectares, the forest coverage rate is 82.0%, and the vegetation is mainly Masson pine, fir, and various broad-leaved tree species. Increased forest coverage reduces sediment production30. The soil is mainly viscous yellow–brown soil and loess parent material. Rain is concentrated in summer, with an average annual rainfall of 1094.6 mm and an average annual temperature of 16.4 °C. Rainfall-related flood risk increased in the Yangtze River Delta in recent years31.The study was based in a Pinus massoniana forest in the Taizishan forest farm of Hubei Province. The Pinus massoniana (Masson pine) is a common species distributed in Central China.Figure 1Geographic location of the study area. Maps were generated using ArcGIS 10.8 for Desktop (http://www.esri.com/software/arcgis/arcgis-for-desktop).Full size imageExperiment designWe chose the Pinus massoniana forest with 47a in the study area as the research object. In the typical Pinus massoniana forest, the separate layers of litter (semi-decomposed and non-decomposed layers) were collected from several 1 m × 1 m quadrat and placed in grid bags. The litter of the semi-decomposed layer have no complete outline, and the color was brown. As the litter leaves of the completely decomposed layer are powdery and are combined with the soil layer, this layer is difficult to collect. Before testing, it was necessary to clean the soil off the pine needles and then allow the litter to dry naturally. The characteristics of the semi-decomposed and non-decomposed litter layers are shown in Table 1. The soil samples need to be dried and screened by 10 mm. When filling the soil trough, every 0.1 m of soil thickness was one layer, for a total of four layers (0.4 m). The characteristics by soil particle sizes are different (Fig. 2). The soil samples were dried naturally, crushed, and then sieved. The soil trough (2 m long, 0.5 m wide and 0.5 m deep) was filled to have a bulk density of 1.53 g·m−3. In this process, an appropriate amount of water was sprinkled on the surface of each soil layer to achieve a soil moisture content consistent with the surrounding, undisturbed, or natural, state. The simulation experiment was conducted in the Jiufeng rainfall laboratory at Beijing Forestry University, China. We used a rainfall simulation system (QYJY-503T, Qingyuan Measurement Technology, Xi’an, China) used a rotary downward spray nozzle. The system is able to simulate a wide range of rainfall intensities (10 to 300 mm h−1) using various water pressure and nozzle sizes controlled by a computer system.Table 1 Characteristics of the non-decomposed and semi-decomposed layers of Pinus massoniana litter.Full size tableFigure 2Soil particle composition of study area soil layers.Full size imageAccording to the results of the field forest investigation, the litter was covered with the experimental treatments shown in Table 2. The treatments mass coverage of non-decomposed litter layer was named as follows: N1 denoted litter mass coverage 0 g·m−2, N2 was ‘the non-decomposed litter mass coverage 100 g·m−2’, N3 was ‘the non-decomposed litter mass coverage 200 g·m−2’, and N4 was ‘the non-decomposed litter mass coverage 400 g·m−2’, N5 was ‘the semi-decomposed litter mass coverage 100 g·m−2’, N6 was ‘the non-decomposed litter mass coverage 100 g·m−2 and the semi-decomposed litter mass coverage 100 g·m−2’, N7 was ‘the non-decomposed litter mass coverage 200 g·m−2 and the semi-decomposed litter mass coverage 100 g·m−2’. N2, N3 and N4 were the undissolved state of litter layer, and N4 (non-decomposed state, ND), N7 (initial stage of litter decomposition, ID), N6 (middle stage of litter decomposition, MD) and N5 (final stage of litter decomposition, FD) respectively represent different stages of litter decomposition.Table 2 The experimental design of this study.Full size tableAccording to the rainfall in the Taizishan area of Hubei Province, erosive rainfall and extreme rainstorms were selected as the research conditions. Summer rainfall events occur mainly in the summer in this area, and a rainfall intensity of 60 mm·h−1 was the most common erosive rainfall intensity. Under extreme weather conditions, the rainfall intensity can reach up to 120 mm·h−1. Our experiments were conducted with 60 and 120 mm·h−1 rain intensities with a rainfall that lasted 1 h. According to the field investigation data of forest land, this area is a low mountain and hilly area with a slope mostly between 5° and 10°. Therefore, 5° and 10° were selected for the slope treatments in this study. The combination of slope and rainfall intensity was named as follows: T1 denoted ‘Slope 5° and rainfall intensity 60 mm·h−1’, T2 was ‘Slope 10° and rainfall intensity 60 mm·h−1’, T3 was ‘Slope 5° and rainfall intensity 120 mm·h−1’, and T4 was ‘Slope 10° and rainfall intensity 120 mm·h−1’. With two rainfall intensities, two slopes, seven litter coverage gradient and two repetitions combined, this study had a total of 56 rainfall events.Experimental procedureBefore the test, the soil samples were wetted for 10 h and then drained for 2 h to eliminate the effect of the initial soil moisture on the soil detachment measurement. When the simulated rainfall started, all the runoff and sediment produced from plot were collected every 5 min in the first 10 min, and then collected once every 10 min during the subsequent 50 min. At the same time, runoff velocity, depth and temperature were measured and vernier calliper (accuracy 0.02 mm) respectively.The overland flow velocity was measured using dying method (KMnO4 solution)32. After judging the flow pattern, we confirmed the correction coefficient K value (in laminar flow state, K = 0.67; transition flow state, K = 0.70; turbulent flow state, K = 0.8). The average velocity of overland flow was obtained by multiplying the correction coefficient K and the instantaneous velocity. Runoff depth was measured using vernier calliper (accuracy 0.02 mm). Runoff temperature was measured using thermometer. When the rainfall experiment finished, the collected runoff samples were measured volumetric cylinder and then settled for at least 12 h. The clear water was decanted, and the samples were put into an oven to dry for 24 h under 105 °C. The sediment sample was dried and weighed with an electronic scale.Calculation of hydrodynamic parametersOverland flow has the characteristics of a thin water layer, large fluctuations of the underlying surface, and unstable flow velocity. At present, most scholars use open-channel flow theory to study overland flow33,34. In open-channel flow theory, the Reynold’s number (Re), Froude constant (Fr), flow index (m), resistance coefficient (f), and soil separation rate (({D}_{r})) are the basic parameters of overland flow dynamics, through Reynold’s number (Re), Froude constant (Fr), flow index (m) can distinguish flow patterns. Re is calculated as:$$Re=Rcdot V/nu ,$$where Re is the Reynolds number of the water flow, which is dimensionless, and can be used to judge the flow state of overland flow. When Re ≤ 500, the flow pattern is laminar; when 500   5000, the flow pattern is turbulent. R is the hydraulic radius (m), which is generally replaced by flow depth as measured by a vernier calliper (accuracy 0.02 mm). (V) is the average velocity (m·s−1); (nu) is the kinematic viscosity coefficient (m2·s−1), and the calculation formula is (nu) = 0.01775·10−4·(1 + 0.0337 t + 0.00021 t2), where t is the test overland flow temperature35.Fr is the Froude constant, which is the ratio of the inertial force to gravity and can be used to distinguish overland flow as rapid flow, slow flow, or critical flow. When Fr  1, the fluid is rapid flow.Fr is calculated as:$$Fr=V/sqrt{gcdot R},$$where (Fr) is the Froude constant of the water flow, which is dimensionless; (V) is the average velocity (m·s−1); g is the acceleration of gravity and has a constant value of 9.8 m·s−2; R is a hydraulic radius (m), and is generally replaced by flow depth as measured by a vernier calliper (accuracy 0.02 mm).Regression fitting is made for runoff depth (h) and single width flow (Q). The runoff depth equation for slope is as follows:$$h=k{q}^{m},$$where q is the single width flow (L·m−1·s−1); h is the depth of water on the slope (m); and m is the flow index, which reflects the turbulent characteristics of the flow state. The larger m is, the more energy the flow consumes in the work of resistance. The comprehensive index (k) reflects the characteristics of the underlying surface and the water viscosity of the slope flow. The larger k is, the stronger the surface material of the slope works on the flow.The resistance of overland flow reflects the inhibition effect of different underlying surface conditions on the velocity of overland flow. The Darcy–Weisbach formula is widely used in research because of its two advantages: applicability and dimensionlessness under laminar and turbulent flow conditions36,37.The resistance coefficient (f) is calculated as follows:$$f=8cdot gcdot Rcdot J/{V}^{2},$$where the resistance coefficient f has no dimension; g is the acceleration of gravity and is always 9.8 m·s−2; R is a hydraulic radius (m), generally replaced by flow depth measured by a vernier calliper (accuracy 0.02 mm); (V) is the average velocity (m·s−1); and J is the hydraulic gradient, which can be converted by the gradient in a uniform flow state and is generally replaced by the sine value of the gradient.Shear stress ((tau)) is the main driving force that affects the stripping of soil particles from the surface soil38. Shear stress is calculated as:$$tau =rcdot gcdot Rcdot J,$$where (tau) is the shear force of runoff (Pa); and r is the density of water and sediment concentration flow (kg·m−3). This study used a muddy water mass and volume ratio in the unseparated state to calculate the density of water and sediment concentration flow.Flow power (W) is the runoff power per unit area of water and refers to the power consumed by the weight of water acting on the riverbed surface to transport runoff and sediment. W is calculated as:$$W=tau cdot V,$$where W is the flow power (N·m−1·s−1); and (tau) is the shear force of runoff (Pa).Soil separation rate (({D}_{r})) refers to the quality of soil in which soil particles are separated from the soil per unit time. The calculation formula is as follows:$${D}_{r}={W}_{d}-{W}_{w}/tcdot A,$$where ({D}_{r}) is the rate of soil separation (kg·m−2·s−1); ({W}_{w}) is the dry weight of soil before the test; ({W}_{d}) is the dry weight of soil after the test, measured by the drying method (kg); t is the scouring time (s); and A is the surface area of the soil sample (m2). More