Transferability and model evaluation in ecological niche modeling: a comparison of GARP and Maxent

Ecography - Tập 30 Số 4 - Trang 550-560 - 2007
A. Townsend Peterson1, Monica Papeş, Muir D. Eaton
1University of Kansas

Tóm tắt

We compared predictive success in two common algorithms for modeling species’ ecological niches, GARP and Maxent, in a situation that challenged the algorithms to be general – that is, to be able to predict the species’ distributions in broad unsampled regions, here termed transferability. The results were strikingly different between the two algorithms – Maxent models reconstructed the overall distributions of the species at low thresholds, but higher predictive levels of Maxent predictions reflected overfitting to the input data; GARP models, on the other hand, succeeded in anticipating most of the species’ distributional potential, at the cost of increased (apparent, at least) commission error. Receiver operating characteristic (ROC) tests were weak in discerning models able to predict into broad unsampled areas from those that were not. Such transferability is clearly a novel challenge for modeling algorithms, and requires different qualities than does predicting within densely sampled landscapes – in this case, Maxent was transferable only at very low thresholds, and biases and gaps in input data may frequently affect results based on higher Maxent thresholds, requiring careful interpretation of model results.

Từ khóa


Tài liệu tham khảo

10.1016/S0304-3800(02)00349-6

Anon, 2001, HYDRO1k Elevation Derivative Database

10.1126/science.1131758

10.1111/j.1365-2486.2005.01000.x

10.1111/j.1466-822X.2004.00128.x

10.1146/annurev.es.16.110185.000351

10.1023/A:1021346712677

10.2307/2531595

Elith J., 2002, Predicting species occurrences: issues of scale and accuracy, 303

10.1111/j.2006.0906-7590.04596.x

Eng J., 2005, ROC analysis: web‐based calculator for ROC curves

10.1017/S0376892997000088

10.1080/02705060.1997.9663548

10.1016/j.tree.2004.07.006

10.1139/f03-120

10.1016/S0304-3800(00)00354-9

Guisan A., 2007, What matters for predicting spatial distributions of trees: techniques, data, or species’ characteristics?

10.1111/j.1472-4642.2007.00342.x

Hijmans R. J., 2005, WorldClim, ver. 1.3

10.1111/j.1654-1103.2002.tb02036.x

10.1111/j.0906-7590.2005.03957.x

10.1046/j.1365-2664.1999.00440.x

10.1016/S0304-3800(99)00113-1

10.1111/j.1365-2699.2006.01460.x

10.1111/j.1365-2699.2006.01594.x

Pereira R. S., 2002, Desktop GARP

10.17161/bi.v3i0.29

10.1016/S0304-3800(99)00023-X

10.1016/S0020-7519(03)00094-8

Phillips S. J., 2006, Maxent software for species habitat modeling, ver. .2.0, 2

Phillips S. J., 2004, Proc. 21st Int. Conf. on Machine Learning, 655

10.1016/j.ecolmodel.2005.03.026

10.1111/j.1365-2699.2006.01466.x

Sauer J. R., 2001, The North American Breeding Bird Survey, results and analysis 1966–2000, ver. 2001.2

Schröder B., 1999, Are habitat models transferable in space and time?, Zeitschr. Ökol. Natursch., 8, 195

10.1098/rstb.2003.1439

10.17161/bi.v2i0.4

10.1016/0378-4754(92)90126-2

Stockwell D. R. B., 2002, Predicting species occurrences: issues of scale and accuracy, 537

10.1016/S0304-3800(01)00388-X

10.1038/nature02121

10.1002/rrr.3450080307

10.1046/j.0305-0270.2003.00991.x

10.1073/pnas.0409902102

Thông tin
Thông tin xuất bản

Ecography

Tập 30 Số 4

550-560

Thông tin tác giả