Abstract
Wolves have returned to Germany since 2000. Numbers have grown to 209 territorial pairs in 2021. XGBoost machine learning, combined with SHAP analysis is applied to predict German wolf pair presence in 2022 for 10 × 10 km grid cells. Model input consisted of 38 variables from open sources, covering the period 2000 to 2021. The XGBoost model predicted well, with 0.91 as the AUC. SHAP analysis ranked the variables: distance to the closest neighboring wolf pair was the main driver for a grid cell to become occupied by a wolf pair. The clustering tendency of related wolves seems to be an important explanatory factor here. Second was the percentage of wooded area. The next eight variables related to wolf presence in the preceding year, except at fifth, eighth and tenth position in the total order: human density (square root) in the grid, percentage arable land and road density respectively. Other variables including the occurrence of wild prey were the weakest predictors. The SHAP analysis also provided crucial added value in identifying a variable that had threshold values where its contribution to the prediction changed from positive to negative or vice versa. For instance, low density of people increased the probability of wolf pair presence, whereas a high density decreased this probability. Cumulative lift techniques showed that the model performed almost four times better than random prediction. The combination of XGBoost, SHAP and cumulative lift techniques is new in wolf management and conservation, allowing for the focusing of educational and financial resources.
Similar content being viewed by others
References
Ansorge H, Kluth G, Hahne S (2006) Feeding ecology of wolves Canis lupus returning to Germany. Acta Theriol (Warsz) 51:99–106. https://doi.org/10.1007/BF03192661
Bassi E, Willis SG, Passilongo D et al. (2015) Predicting the spatial distribution of wolf () breeding areas in a mountainous region of central Italy. PLoS One 10:e0124698. https://doi.org/10.1371/journal.pone.0124698
Bergstra J, Komer B, Eliasmith C et al. (2015) Hyperopt: a Python library for model selection and hyperparameter optimization. Comput Sci Discov 8:014008. https://doi.org/10.1088/1749-4699/8/1/014008
Bessa-Gomes C, Petrucci-Fonseca F (2003) Using artificial neural networks to assess wolf distribution patterns in Portugal. Anim Conserv 6:221–229. https://doi.org/10.1017/S1367943003003275
Blanco JC, Reig S, de la Cuesta L (1992) Distribution, status and conservation problems of the wolf Canis lupus in Spain. Biol Conserv 60:73–80. https://doi.org/10.1016/0006-3207(92)91157-N
Burbaitė L, Csányi S (2009) Roe deer population and harvest changes in Europe. Estonian J Ecol 58:169. https://doi.org/10.3176/eco.2009.3.02
Caniglia R, Fabbri E, Galaverni M et al. (2014) Noninvasive sampling and genetic variability, pack structure, and dynamics in an expanding wolf population. J Mammal 95:41–59. https://doi.org/10.1644/13-MAMM-A-039
Chapron G, Kaczensky P, Linnell JDC et al. (2014) Recovery of large carnivores in Europe’s modern human-dominated landscapes. Science (1979) 346:1517–1519. https://doi.org/10.1126/science.1257553
Chen T, Guestrin C (2016) XGBoost. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. ACM, New York, NY, USA, pp 785–794
Cimatti M, Ranc N, Benítez‐López A et al. (2021) Large carnivore expansion in Europe is associated with human population density and land cover changes. Divers Distrib 27:602–617. https://doi.org/10.1111/ddi.13219
DBBW (2023) DBBW, the Federal Documentation and Consultation Centre on Wolves. https://www.dbb-wolf.de/home. Accessed 15 May 2023
DIVA-GIS (2011) DIVA-GIS. http://www.diva-gis.org/gdata. Accessed 27 Jul 2023
Eggermann J, da Costa GF, Guerra AM et al. (2011) Presence of Iberian wolf (Canis lupus signatus) in relation to land cover, livestock and human influence in Portugal. Mamm Biol 76:217–221
Erdas C (2020) Wolves and Ravens: Defining a unique relationship. Osmosis Magazine
European Environment Agency (2021) Natura 2000 data - the European network of protected sites. In: https://www.eea.europa.eu/en/datahub/datahubitem-view/6fc8ad2d-195d-40f4-bdec-576e7d1268e4
EUROSTAT (2020) GISCO: GEOGRAPHICAL INFORMATION AND MAPS. In: https://ec.europa.eu/eurostat/web/gisco/geodata/reference-data/administrative-units-statistical-units/countries
Falcucci A, Maiorano L, Tempio G et al. (2013) Modeling the potential distribution for a range-expanding species: Wolf recolonization of the Alpine range. Biol Conserv 158:63–72. https://doi.org/10.1016/j.biocon.2012.08.029
Gantchoff MG, Beyer DE, Erb JD et al. (2022) Distribution model transferability for a wide-ranging species, the Gray Wolf. Sci Rep. 12:13556. https://doi.org/10.1038/s41598-022-16121-6
Gazzola A, Bertelli I, Avanzinelli E et al. (2005) Predation by wolves (Canis lupus) on wild and domestic ungulates of the western Alps, Italy. J Zool 266:205–213. https://doi.org/10.1017/S0952836905006801
GBIF (2023) GBIF, the Global Biodiversity Information Facility. In: https://www.gbif.org/occurrence/download/0224656-230224095556074
Glenz C, Massolo A, Kuonen D, Schlaepfer R (2001) A wolf habitat suitability prediction study in Valais (Switzerland). Landsc Urban Plan 55:55–65. https://doi.org/10.1016/S0169-2046(01)00119-0
Gouwy J, Van Den Berge K, Berlengee F, Mergeay J (2019) Wolvenspecial Oktober 2019. Roofdiernieuws Oktober:1–8
Grilo C, Lucas PM, Fernández‐Gil A et al. (2019) Refuge as major habitat driver for wolf presence in human‐modified landscapes. Anim Conserv 22:59–71. https://doi.org/10.1111/acv.12435
Halvorsen R (2013) A strict maximum likelihood explanation of MaxEnt, and some implications for distribution modelling. Sommerfeltia 36:1–132. https://doi.org/10.2478/v10208-011-0016-2
HDX (2019) Germany: High Resolution Population Density Maps + Demographic Estimates. In: https://data.humdata.org/dataset/germany-high-resolution-population-density-maps-demographic-estimates
Hyperopt (2023) hyperopt package 0.2.7. In: https://pypi.org/project/hyperopt/
Jansman HAH, Mergeay J, Van Der Grift EA, et al. (2021) De wolf terug in Nederland: een factfinding study. Wageningen
Jarausch A, Harms V, Kluth G et al. (2021) How the west was won: genetic reconstruction of rapid wolf recolonization into Germany’s anthropogenic landscapes. Heredity (Edinb) 127:92–106. https://doi.org/10.1038/s41437-021-00429-6
Jȩdrzejewski WŁ, Jȩdrzejewska B, Okarma H et al. (2000) Prey selection and predation by wolves in Bialowieza primeval forest, Poland. J Mammal 81:197–212. https://doi.org/10.1644/1545-1542(2000)081<0197:PSAPBW>2.0.CO;2
Khorozyan I, Heurich M (2022) Large-scale sheep losses to wolves (Canis lupus) in Germany are related to the expansion of the wolf population but not to increasing wolf numbers. Front Ecol Evol 10: https://doi.org/10.3389/fevo.2022.778917
Kittle AM, Anderson M, Avgar T, et al. (2017) Landscape‐level wolf space use is correlated with prey abundance, ease of mobility, and the distribution of prey habitat. Ecosphere 8: https://doi.org/10.1002/ecs2.1783
Kramer-Schadt S, Wenzler M, Gras P, Knauer F (2020) Habitatmodellierung und Abschätzung der potenziellen Anzahl von Wolfsterritorien in Deutschland. Deutschland/Bundesamt für Naturschutz
Kuhn M, Johnson K (2013) Applied Predictive Modeling. Springer New York, New York, NY
Kuijper DPJ, Sahlén E, Elmhagen B, et al. (2016) Paws without claws? Ecological effects of large carnivores in anthropogenic landscapes. Proceedings of the Royal Society B: Biological Sciences 283: https://doi.org/10.1098/rspb.2016.1625
Lundberg SM, Erion G, Chen H et al. (2020) From local explanations to global understanding with explainable AI for trees. Nat Mach Intell 2:56–67. https://doi.org/10.1038/s42256-019-0138-9
Lundberg SM, Lee SI (2017) A unified approach to interpreting model predictions. In: Advances in neural information processing systems, 30. 4765–4774
Marucco F, Boiani MV, Dupont P et al. (2023) A multidisciplinary approach to estimating wolf population size for long‐term conservation. Conservation Biology. https://doi.org/10.1111/cobi.14132
Massolo A, Meriggi A (1998) Factors affecting habitat occupancy by wolves in northern Apennines (northern Italy): a model of habitat suitability. Ecography 21:97–107. https://doi.org/10.1111/j.1600-0587.1998.tb00663.x
Mayer M, Olsen K, Schulz B et al. (2022) Occurrence and livestock depredation patterns by wolves in highly cultivated landscapes. Front Ecol Evol 10: https://doi.org/10.3389/fevo.2022.783027
Mech LD, Boitani L (2003) Wolves: Behavior, Ecology, and Conservation. The University of Chicago Press, Chicago, IL, USA
Mech LD, Dawson DK, Peek JM et al. (1980) Deer Distribution in Relation to Wolf Pack Territory Edges. J Wildl Manag 44:253. https://doi.org/10.2307/3808381
Mech LD, Harper EK (2002) Differential use of a wolf, Canis lupus, pack territory edge and core. Can Field-Naturalist 116:315–316
Mladenoff DJ, Clayton MK, Pratt SD, et al. (2009) Change in Occupied Wolf Habitat in the Northern Great Lakes Region. In: Recovery of Gray Wolves in the Great Lakes Region of the United States. Springer New York, New York, NY, pp 119–138
Mladenoff DJ, Sickley TA, Sickley TA et al. (1995) A regional landscape analysis and prediction of favorable gray wolf habitat in the northern Great Lakes region. Conserv Biol 9:279–294
Modelplotpy (2023) modelplotpy package, 1.0.0. In: https://modelplotpy.readthedocs.io/en/latest/
Nagelkerke J (2022) Visualise the business value of predictive models. In: https://medium.com/p/21c6bc8132c
Naturhistorisk museum Aarhus (2023) Atlas over Danmarks Ulve. In: https://www.ulveatlas.dk/nyheder/ulvehvalpe-i-danmark-for-foerste-gang-i-over-200-aar/
Nowak S, Mysłajek RW, Kłosińska A, Gabryś G (2011) Diet and prey selection of wolves (Canis lupus) recolonising Western and Central Poland. Mamm Biol 76:709–715. https://doi.org/10.1016/j.mambio.2011.06.007
Oakleaf JK, Murray DL, Oakleaf JR et al. (2006) Habitat selection by recolonizing wolves in the northern Rocky Mountains of the United States. J Wildl Manag 70:554–563
Okarma H (1995) The trophic ecology of wolves and their predatory role in ungulate communities of forest ecosystems in Europe. Acta Theriol (Warsz) 40:335–386. https://doi.org/10.4098/AT.arch.95-35
Ordiz A, Uzal A, Milleret C et al. (2020) Wolf habitat selection when sympatric or allopatric with brown bears in Scandinavia. Sci Rep. 10:9941. https://doi.org/10.1038/s41598-020-66626-1
Pasini A (2015) Artificial neural networks for small dataset analysis. J Thorac Dis 7:953–960
Pedregosa F, Varoquaux G, Gramfort A et al. (2011) Scikit-learn: Machine Learning in Python. J Mach Learn Res 12:2825–2830
Reinhardt I, Ansorge H, Collet S et al. (2021) Erkenntnisse zur Wiederausbreitung des Wolfs in Deutschland. 0028-0615 96:19–26. https://doi.org/10.17433/1.2021.50153869.19-26
Reinhardt I, Kluth G (2015) Untersuchungen zum Raum-Zeitverhalten und zur Abwanderung von Wölfen in Sachsen. Endbericht Projekt “Wanderwolf” (2012 - 2014)
Reinhardt I, Kluth G, Jarausch A et al. (2017) Dokumentations- und Beratungsstelle des Bundes zum Thema Wolf. Wölfe in Deutschland - Statusbericht 2015/16.
Reinhardt I, Kluth G, Nowak C et al. (2019) Military training areas facilitate the recolonization of wolves in Germany. Conserv Lett 12:. https://doi.org/10.1111/conl.12635
Ripple WJ, Estes JA, Beschta RL et al. (2014) Status and Ecological Effects of the World’s Largest Carnivores. Science (1979) 343:. https://doi.org/10.1126/science.1241484
Roche DG, O’Dea RE, Kerr KA, et al. (2022) Closing the knowledge‐action gap in conservation with open science. Conservation Biology 36:. https://doi.org/10.1111/cobi.13835
Roder S, Biollaz F, Mettaz S et al. (2020) Deer density drives habitat use of establishing wolves in the Western European Alps. J Appl Ecol 57:995–1008. https://doi.org/10.1111/1365-2664.13609
Scikit-learn (2023) scikit-learn package 1.2.2. In: https://pypi.org/project/scikit-learn/
SHAP (2023) shap 0.41.0. In: https://pypi.org/project/shap/
Smith JB, Nielsen CK, Hellgren EC (2016) Suitable habitat for recolonizing large carnivores in the midwestern USA. Oryx 50:555–564. https://doi.org/10.1017/S0030605314001227
Stahler D, Heinrich B, Smith D (2002) Common ravens, Corvus corax, preferentially associate with grey wolves, Canis lupus, as a foraging strategy in winter. Anim Behav 64:283–290. https://doi.org/10.1006/anbe.2002.3047
Tuia D, Kellenberger B, Beery S et al. (2022) Perspectives in machine learning for wildlife conservation. Nat Commun 13:792. https://doi.org/10.1038/s41467-022-27980-y
van den Bosch M, Beyer DE, Erb JD et al. (2022) Identifying potential gray wolf habitat and connectivity in the eastern USA. Biol Conserv 273:109708. https://doi.org/10.1016/j.biocon.2022.109708
Van der Veken T, Van den Berge K, Gouwy J et al. (2021) Diet of the first settled wolves (Canis lupus) in Flanders, Belgium. Lutra 64:45–56
van Liere D, Siard N, Martens P, Jordan D (2021) Conflicts with wolves can originate from their parent packs. Animals 11:1801. https://doi.org/10.3390/ani11061801
Wagner C, Holzapfel M, Kluth G et al. (2012) Wolf (Canis lupus) feeding habits during the first eight years of its occurrence in Germany. Mamm Biol 77:196–203. https://doi.org/10.1016/j.mambio.2011.12.004
Witek M, Zwolicki A, Wikar Z et al. (2023) High abundance of an introduced prey species, fallow deer Dama dama, abolishes wolf preference towards red deer. In: Wolves across borders, international conference on wolf ecology and management
XGBoost (2023) xgboost 1.7.5. In: https://pypi.org/project/xgboost/
Zabihi-Seissan S, Prokopenko CM, Vander Wal E (2022) Wolf spatial behavior promotes encounters and kills of abundant prey. Oecologia 200:11–22. https://doi.org/10.1007/s00442-022-05218-4
Acknowledgements
First and foremost, we would like to express our gratitude to the members of our research team, Simone Spierings, Kjeld Vissers, Philip Vermeij and Erik van Nistelrooij, who provided valuable input, insights, and assistance at every stage of the project. Their contributions were critical to the success of this research, and we are deeply grateful for their hard work and dedication. We also thank dr Markus Ritz for allowing us to use the DBBW (Dokumentations- und Beratungsstelle des Bundes zum Wolf) data, that are provided to the DBBW by the German federal states.
Author information
Authors and Affiliations
Contributions
JS: Methodology, Validation, Formal analysis, Writing—Original draft, Visualization, Project administration. JN: Methodology, Validation, Formal analysis, Writing—Review & editing. TS: Formal analysis, Writing—Review & editing. NO: Conceptualization, Writing—Review & editing. DvanL: Conceptualization, Writing—Original draft All authors reviewed the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
The authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Schoonemann, J., Nagelkerke, J., Seuntjens, T.G. et al. Applying XGBoost and SHAP to Open Source Data to Identify Key Drivers and Predict Likelihood of Wolf Pair Presence. Environmental Management 73, 1072–1087 (2024). https://doi.org/10.1007/s00267-024-01941-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00267-024-01941-1