The influence of habitat patches and habitat qualities on small rodent occupancy patterns in Joshua Tree National Park

Background/Question/Methods

Naturally occurring events such as glaciations, fires, and floods have frequently disrupted continuous expanses of natural habitat. However, dramatic increases in global human population growth coupled with urbanization, agriculture, and deforestation have rapidly accelerated the rate of habitat fragmentation, resulting in widespread changes in the spatial structure of native habitats. Furthermore, climate change has indirectly altered key disturbance regimes creating an increase in non-natural habitat fragmentation. Understanding how the landscape influences community dynamics in a spatially heterogeneous landscape is critical in order to explain long-term effects of habitat fragmentation on population persistence.

In Joshua Tree National Park (JTNP), wildfire was infrequent prior to recent times. However, fires are now increasingly common due to the invasion of non-native grass species which has been facilitated by climate change and an on-going nitrogen soil deposition from urban California. This altered fire regime has left a patchwork of burned and unburned habitat patches which creates a natural experiment to test the influence of habitat heterogeneity on community patterns within a fragmented landscape. We measured small rodent occupancy patterns in five burned macroplots varying in habitat quality and isolation levels, 12 remnant habitat patches, and one nearby, unburned control macroplot. 

Results/Conclusions

We ran a Kruskal-Wallis rank sum non-parametric test in order to determine if there are significant differences in rodent occupancy patterns among habitat qualities. When there are remnant habitat patches present within a burned habitat, there is no significant difference between the burned habitat and the unburned control for rodent occupancy patterns, even at relatively isolated distances from unburned habitat. There is also no significant difference in occupancy patterns between the unburned habitat and the burned habitat that is relatively close to unburned habitat. However, when a burned habitat does not have remnant habitat patches and is isolated from the unburned habitat, it has significantly lower rodent occupancy patterns. Furthermore, when we separate remnant habitat patches into two categories:  clustered patches (flower-like arrangement) or linear designs, we see that clustered patches maintain significantly higher rodent abundances than patches in a linear arrangement.

Our results suggest that remnant habitat patches can mask the effect of isolation levels in a burned habitat. Furthermore, habitat patches in a clustered arrangement will maintain higher rodent abundances than patches in a linear design. For future restoration techniques, our study emphasizes the importance of vegetation restoration plantings after novel disturbances in order to promote recolonization.