Fire and grazing regulate the reptile and amphibian community patch dynamics in tallgrass prairie

Background/Question/Methods

Fire and cattle grazing are common management schemes of tallgrass prairie, and are potentially important drivers of reptilian and amphibian (herpetofauna) metapopulation dynamics.  Few studies have assessed the impacts of fire, grazing, the interaction, and recovery phases to herpetofauna in any ecosystem; this is especially true in tallgrass prairie. A patch-burn grazing study at Osage Prairie in MO created landscape patches of (1) Control (no fire or grazing in the last 5 years), (2) Fire, (3) Lt Grazing (light cattle grazing, no fire), (4) F+G (fire followed by heavy cattle grazing), (5) Fire-Legacy (burned previous year), and (6) F+G-Legacy (burned and grazed the previous year). This landscape mosaic created an opportunity to study these management techniques on the patch dynamics of riparian herpetofauna. I used robust-design occupancy modeling to measure patch occupancy, detection, colonization and extinction in each patch type during six primary sampling seasons in the spring of 2011 and 2012 using a variety of sampling techniques. To determine if treatment and environmental factors were important in explaining community assemblage, I conducted redundancy analysis (RDA) and a multivariate analysis of variance (MANOVA) for both amphibians and reptiles. Further, I calculated Shannon’s diversity index for each patch and each watershed. 

Results/Conclusions

Detection rates were only ~50% (p ~ 0.5), suggesting occupancy modeling is a necessary analytical approach to obtain accurate parameter estimates. Estimates for reptiles indicate occupancy was seasonally constant in control patches (ψ ~ 0.5), but declined to ψ ~ 0.2 in fire and grazed patches over time as these treatments were applied. Extinction rates for reptiles were greatest in the fire and light grazed plots following treatment (ϵ ~ 0.7), and colonization was negligible in all patches. For reptiles, the RDA (p<0.001) and MANOVA found patch type (p<0.001) and grass height (p=0.048) as important predictors of abundance. Interestingly, turtles were strongly associated with the patch Lt Grazing and tall grasses, snakes with the Fire-Legacy patch and high percentage of vegetation cover, and lizards with Fire, F+G, and F+G-Legacy patch types and bare soil. The amphibian RDA (p<0.001) and MANOVA discovered year (p<0.001), site (p=0.002), in-stream cover (p=0.007), and substrate type (p=0.029) as predictors, but not patch type (p=0.101). The varying responses by taxonomic grouping to the patch types provide support for patch-burning grazing management if the goal is increasing landscape diversity.