Stochastic dispersal shapes the spatial pattern of species richness in mountain landscapes

<p class="MsoNormal"><strong><span>Aim<a name="OLE_LINK3"></a>: </span></strong><span><span>Biogeographers have begun to address the problem of species distribution patterns in three-dimensional space. A key question is: What patterns of species richness would arise on the three-dimensional surface of a landscape under minimal biological assumptions? Recently, a theory called "Landscape Elevational Connectivity" (LEC) has been developed, which measures how topography and geomorphology drive biodiversity patterns. Here, we tested the predictive ability of LEC for spatial patterns of species richness for the first time.</span></span></p> <p class="MsoNormal"><span><strong><span>Location: </span></strong></span><span><span>The Tibetan Plateau.</span></span></p> <p class="MsoNormal"><span><strong><span>Methods:</span></strong></span><span><span> We </span><span>used the "stacked species distribution models" (S-SDMs) approach to</span></span><span><span> estimate the empirical spatial distribution pattern of bird species richness on the Tibetan Plateau based on online species occurrence data and expert maps, and we compared this estimated distribution with the predictions of LEC.</span></span></p> <p class="MsoNormal"><span><strong><span>Results: </span></strong></span><span><span>We found a high correlation between the LEC null model and observed bird species richness in the biodiversity hotspot on the southeast edge of the Tibetan Plateau (Spearman's correlation, <em>r</em><span>s</span> = 0.746, 95% CI: 0.744-0.748). On a wider scale, LEC was better correlated with species richness in regions higher net primary productivity than in regions with lower net primary productivity.</span></span></p> <p class="MsoNormal"><span><strong><span>Main conclusions:</span></strong></span><span><span> <a name="OLE_LINK31"></a>Our results suggest that the impact of stochastic processes on the spatial distribution pattern of species richness may have been routinely underestimated, especially in regions with rich resources and high species richness. We conclude that it would be fruitful to reconsider the contribution of deterministic factors to the distribution pattern of species richness, especially in mountain landscapes, by applying LEC as a null model.</span></span></p>