Informative incongruences between niche based models and process based models

Background/Question/Methods

Niche-based models (NBM) are based on the empirical statistical modeling approaches that correlate environmental variables to species’ presence and / or absence to determine its distribution. NBM profile the bioclimatic envelope of the species (realized niche) through a number of different statistical techniques in order to project them to the geographical space. This correlative approach has extensively been used to address many challenges posed by climate change on species distributions such as extinctions, species range change or invasive species. However, direct processes are not being examined and pseudo-equilibrium with climate is assumed. 

On the other hand, process-based models (PBMs) represent a modeling approach that focuses on the mechanisms that directly or indirectly shape species distributions (e.g. growth, migration, reproduction, etc), based on physical laws. The agreement between patterns (NBMs) and processes (PBMs) may help in assessing to what extent climate variables or other factors are shaping current distributions at a given scale.

An ensemble of NBMs was built to project suitability values for three common Mediterranean tree species (Pinus halepensis, Pinus sylvestris and Quercus ilex) in the Iberian Peninsula. Using GOTILWA+ model, growth for these species was simulated along the previously modeled suitability values.  Outputs from these two models were then compared using a robust regression.

Results/Conclusions

Results showed that for two of the three tree species (Pinus sylvestris and Quercus ilex) growth and suitability were in agreement, presenting a positive relationship between the two outputs.  However, some outliers may be identified where forests may potentially grow under a low value of suitability. For one species (Pinus halepensis), growth and suitability presented a negative trend showing no congruence.

Mapping models’ agreement and disagreement regions enables to geographically assess the influence of climate on species distributions. Regions were models are incongruent represent the geographical space where other processes (e.g: dispersal, competition, etc) may be playing a key role. Agreement regions represent areas where NBM suitability values may be reliable to use in subsequent studies.  We emphasize that model comparison is necessary to understand current modeling tools as well as refining our hypothesis regarding species distribution.