Selecting thresholds of occurrence in the prediction of species distributions

Ecography - Tập 28 Số 3 - Trang 385-393 - 2005
Canran Liu1, Pam Berry2,3, Terence P. Dawson2,3, Richard G. Pearson2,3
1University of New England
2Dept of Ecosystem Management, School of Environmental Sciences & Natural Resources Management, Univ. of New England, Armidale, NSW 2351, Australia.
3Environmental Change Inst., Centre for the Environment, Univ. of Oxford, Dyson Perrins Building, South Parks Road, Oxford, UK OX1 3QY.

Tóm tắt

Transforming the results of species distribution modelling from probabilities of or suitabilities for species occurrence to presences/absences needs a specific threshold. Even though there are many approaches to determining thresholds, there is no comparative study. In this paper, twelve approaches were compared using two species in Europe and artificial neural networks, and the modelling results were assessed using four indices: sensitivity, specificity, overall prediction success and Cohen's kappa statistic. The results show that prevalence approach, average predicted probability/suitability approach, and three sensitivity‐specificity‐combined approaches, including sensitivity‐specificity sum maximization approach, sensitivity‐specificity equality approach and the approach based on the shortest distance to the top‐left corner (0,1) in ROC plot, are the good ones. The commonly used kappa maximization approach is not as good as the afore‐mentioned ones, and the fixed threshold approach is the worst one. We also recommend using datasets with prevalence of 50% to build models if possible since most optimization criteria might be satisfied or nearly satisfied at the same time, and therefore it's easier to find optimal thresholds in this situation.

Từ khóa


Tài liệu tham khảo

10.1016/S0304-3800(02)00205-3

10.1016/S0006-3207(02)00119-2

10.1111/j.0906-7590.2004.03553.x

Berry P. M., 2001, Climate change and nature conservation in the Britain and Ireland: modelling natural resource responses to climate change (the MONARCH project), 43

Berry P. M., 2002, Modelling potential impacts of climate change on the bioclimatic envelope of species in Britain and Ireland, Global Ecol. Biogeogr., 11, 453, 10.1111/j.1466-8238.2002.00304.x

10.1016/S0304-3800(99)00110-6

10.1016/S0895-4356(99)00075-X

10.1046/j.1365-2664.2000.00556.x

10.1017/CBO9780511615412

10.1046/j.1365-2699.2000.00419.x

10.1016/S0304-3800(02)00260-0

Fielding A. H., 1995, Testing the generality of bird‐habitat models, Conserv. Biol., 9, 1446, 10.1046/j.1523-1739.1995.09061466.x

10.1017/S0376892997000088

10.1127/phyto/30/2000/353

10.1016/S0304-3800(00)00354-9

10.2307/3237224

10.1023/A:1009841519580

10.1016/S0304-3800(02)00204-1

10.1016/S0304-3800(01)00323-4

10.2307/2845830

10.1023/A:1009754417131

10.1016/S0006-3207(01)00223-3

10.1016/S0304-3800(99)00113-1

10.1046/j.1365-2664.2001.00647.x

10.1016/S0001-2998(78)80014-2

10.1016/S0304-3800(02)00197-7

Nahm U. Y., 2000, Using information extraction to aid the discovery of prediction rules from text, 51

10.1046/j.1523-1739.2003.01280.x

10.1577/1548-8659(2001)130<0878:FHRILG>2.0.CO;2

10.1046/j.1365-2664.2001.00604.x

10.1016/S0304-3800(98)00149-5

10.1016/S0304-3800(00)00322-7

10.1016/S0304-3800(02)00056-X

10.1111/j.0906-7590.2004.03740.x

10.1046/j.1472-4642.2001.00094.x

10.1016/S0304-3800(03)00028-0

Schapire R. E., 1998, Boosting and Rocchio applied to text filtering, 215

10.1016/S0304-3800(01)00388-X

10.1016/S0006-3207(01)00164-1

10.1093/clinchem/39.4.561