Why finer spatial resolutions do not necessarily improve species distribution models
Abstract
In recent years, there has been growing criticism of using too coarse environmental data in species distribution models (SDMs), especially for sessile organisms such as plants. However, we argue that coarse‐resolution environmental predictors are not only conceptually in line with the idea of SDMs, but may often also lead to more robust and realistic models than those with resolutions approaching the sizes of individuals. We explicitly demonstrate this seeming paradox using a virtual ecologist experiment. The main reason behind this is the disequilibrium between a species' fundamental niche and its realised distribution. While this disequilibrium is well‐recognised in biogeography and metapopulation ecology, its impact on fine‐scale SDMs remains rather neglected. Contrary to recent claims, we therefore suggest that the optimal spatial resolution of predictors in SDMs should be guided less by the size of individuals and rather by the spatial extent required to sustain minimum viable populations. Synthesis: Seeking the perfect spatial resolution should not be regarded as the holy grail of ecological modelling. In reality, no such universally optimal resolution exists. Progress is more likely to come from explicitly integrating processes operating across multiple resolutions. In particular, incorporating neighbourhood‐scale variables as predictors in SDMs represents an important step in this direction.
Related
- Advancing Spatial Pattern Integration in Species Distribution Modelling With Computer Vision: How Grain Size and Extent Concepts Shape Predictions of Emerging Aquatic Insects
- A field observation method to characterize site soundscape and improve detection probability estimation for eastern bat species in North America
- When does rural population decline improve ecological quality? Evidence from population-density thresholds and spatial heterogeneity across Chinese counties
- Distribution dataset of chondrichthyan species occurring in French metropolitan waters and across the North-East Atlantic
- Pine needles as an indicator of the spatial distribution of trace elements in an urban-industrial area
BibTeX
@article{chytry2026finer,
title = {{Why finer spatial resolutions do not necessarily improve species distribution models}},
author = {Kryštof Chytrý and Jonathan Lenoir and Dietmar Moser and Karl Hülber and Norbert Helm and Johannes Hausharter and Johannes Wessely and Stefan Dullinger},
journal = {Journal of Ecology},
year = {2026},
month = oct,
doi = {10.1111/1365-2745.70451},
url = {https://u-picardie.hal.science/hal-05795079v1},
note = {CC BY},
}
RIS
TY - JOUR
TI - Why finer spatial resolutions do not necessarily improve species distribution models
AU - Kryštof Chytrý
AU - Jonathan Lenoir
AU - Dietmar Moser
AU - Karl Hülber
AU - Norbert Helm
AU - Johannes Hausharter
AU - Johannes Wessely
AU - Stefan Dullinger
PY - 2026
DA - 2026/10/07
JO - Journal of Ecology
DO - 10.1111/1365-2745.70451
UR - https://u-picardie.hal.science/hal-05795079v1
AB - In recent years, there has been growing criticism of using too coarse environmental data in species distribution models (SDMs), especially for sessile organisms such as plants. However, we argue that coarse‐resolution environmental predictors are not only conceptually in line with the idea of SDMs, but may often also lead to more robust and realistic models than those with resolutions approaching the sizes of individuals. We explicitly demonstrate this seeming paradox using a virtual ecologist experiment. The main reason behind this is the disequilibrium between a species' fundamental niche and its realised distribution. While this disequilibrium is well‐recognised in biogeography and metapopulation ecology, its impact on fine‐scale SDMs remains rather neglected. Contrary to recent claims, we therefore suggest that the optimal spatial resolution of predictors in SDMs should be guided less by the size of individuals and rather by the spatial extent required to sustain minimum viable populations. Synthesis: Seeking the perfect spatial resolution should not be regarded as the holy grail of ecological modelling. In reality, no such universally optimal resolution exists. Progress is more likely to come from explicitly integrating processes operating across multiple resolutions. In particular, incorporating neighbourhood‐scale variables as predictors in SDMs represents an important step in this direction.
SN - 0022-0477
SN - 1365-2745
ER -