A data-integration approach to correct sampling bias in species distribution models using multiple datasets of breeding birds in the Swiss Alps

被引:8
作者
Tehrani, Nasrin Amini [1 ]
Naimi, Babak [2 ]
Jaboyedoff, Michel [1 ]
机构
[1] Univ Lausanne, Fac Geosci & Environm, CH-1015 Lausanne, Switzerland
[2] Univ Helsinki, Dept Geosci & Geog, POB 64, Helsinki 00014, Finland
关键词
Bird species; Data pooling; Model-based data integration; Presence-only data; Random forest; Species distribution models; Swiss Alps; HABITAT SUITABILITY; CITIZEN SCIENCE; SPATIAL SCALES; RANDOM FOREST; CONSERVATION; BIODIVERSITY; ACCURACY; UNCERTAINTY; PREDICTIONS; PERFORMANCE;
D O I
10.1016/j.ecoinf.2021.101501
中图分类号
Q14 [生态学(生物生态学)];
学科分类号
071012 ; 0713 ;
摘要
It is essential to accurately model species distributions and biodiversity in response to many ecological and conservation challenges. The primary means of reliable decision-making on conservation priority are the data on the distributions and abundance of species. However, finding data that is accurate and reliable for predicting species distribution could be challenging. Data could come from different sources, with different designs, coverage, and potential sampling biases. In this study, we examined the emerging methods of modelling species distribution that integrate data from multiple sources such as systematic or standardized and casual or occasional surveys. We applied two modelling approaches, "data-pooling" and " model-based data integration" that each involves combining various datasets to measure environmental interactions and clarify the distribution of spe-cies. Our paper demonstrates a reliable data integration workflow that includes gathering information on model -based data integration, creating a sub-model of each dataset independently, and finally, combining it into a single final model. We have shown that this is a more reliable way of developing a model than a data pooling strategy that combines multiple data sources to fit a single model. Moreover, data integration approaches could improve the poor predictive performance of systematic small datasets, through model-based data integration techniques that enhance the predictive accuracy of Species Distribution Models. We also identified, consistent with previous research, that machine learning algorithms are the most accurate techniques to predict bird species distribution in our heterogeneous study area in the western Swiss Alps. In particular, tree-dependent ensembles of Random Forest (RF) contribute to a better understanding of the interactions between species and the environment.
引用
收藏
页数:9
相关论文
共 101 条
[1]   An assessment of integrated population models: bias, accuracy, and violation of the assumption of independence [J].
Abadi, Fitsum ;
Gimenez, Olivier ;
Arlettaz, Raphael ;
Schaub, Michael .
ECOLOGY, 2010, 91 (01) :7-14
[2]   Fit-for-Purpose: Species Distribution Model Performance Depends on Evaluation Criteria - Dutch Hoverflies as a Case Study [J].
Aguirre-Gutierrez, Jesus ;
Carvalheiro, Luisa G. ;
Polce, Chiara ;
van Loon, E. Emiel ;
Raes, Niels ;
Reemer, Menno ;
Biesmeijer, Jacobus C. .
PLOS ONE, 2013, 8 (05)
[3]   Assessing the accuracy of species distribution models: prevalence, kappa and the true skill statistic (TSS) [J].
Allouche, Omri ;
Tsoar, Asaf ;
Kadmon, Ronen .
JOURNAL OF APPLIED ECOLOGY, 2006, 43 (06) :1223-1232
[4]   Toward community predictions: Multi-scale modelling of mountain breeding birds' habitat suitability, landscape preferences, and environmental drivers [J].
Amini Tehrani, Nasrin ;
Naimi, Babak ;
Jaboyedoff, Michel .
ECOLOGY AND EVOLUTION, 2020, 10 (12) :5544-5557
[5]  
[Anonymous], 2021, IEEE Trans. Broadcast.
[6]   Five (or so) challenges for species distribution modelling [J].
Araujo, Miguel B. ;
Guisan, Antoine .
JOURNAL OF BIOGEOGRAPHY, 2006, 33 (10) :1677-1688
[7]   Selecting pseudo-absences for species distribution models: how, where and how many? [J].
Barbet-Massin, Morgane ;
Jiguet, Frederic ;
Albert, Cecile Helene ;
Thuiller, Wilfried .
METHODS IN ECOLOGY AND EVOLUTION, 2012, 3 (02) :327-338
[8]   Predicting habitat suitability with machine learning models:: The potential area of Pinus sylvestris L. in the Iberian Peninsula [J].
Benito Garzon, Marta ;
Blazek, Radim ;
Neteler, Markus ;
Sanchez de Dios, Rut ;
Sainz Ollero, Helios ;
Furlanello, Cesare .
ECOLOGICAL MODELLING, 2006, 197 (3-4) :383-393
[9]  
Bikkina V., 2014, COMP MACHINE LEARNIN
[10]   Statistical solutions for error and bias in global citizen science datasets [J].
Bird, Tomas J. ;
Bates, Amanda E. ;
Lefcheck, Jonathan S. ;
Hill, Nicole A. ;
Thomson, Russell J. ;
Edgar, Graham J. ;
Stuart-Smith, Rick D. ;
Wotherspoon, Simon ;
Krkosek, Martin ;
Stuart-Smith, Jemina F. ;
Pecl, Gretta T. ;
Barrett, Neville ;
Frusher, Stewart .
BIOLOGICAL CONSERVATION, 2014, 173 :144-154