We recently completed the installation of experimental infrastructure for long-term research studies in Great Basin sagebrush-steppe and Mojave shrubland communities, both of which are currently experiencing dramatic increases in exotic grass invasion and fire. All studies discussed below will be conducted at both the Great Basin and Mojave Desert research experiments using identical experimental design unless otherwise noted. To balance ecological realism with the need for rigorously controlled experiments that can isolate specific mechanisms, we are using both large-scale manipulation studies in the field (Figure below), and a landscape-level network of 6 paired burned and unburned transects (1 km in length) that occur along burn boundaries of multiple independent fires in the Beaver Dam Wash (Mojave only). The large-scale manipulation experiments implement a full-factorial design allowing for an examination of the main and interactive effects of variable precipitation, fire and small mammal abundance, which we hypothesize to be the dominant controls over desert ecosystem function.
Aerial photo of the Great Basin experiment with a full factorial design in 5 replicated blocks. Treatments include: Fire B = burned, U = unburned; Small mammals presence: Y = present, N = excluded using small mammal fences ; Precipitation: r = 30% reduction, s = 30% supplement applied in 3 week intervals, c = ambient control using 2 × 3 mrainout shelters (upper right) designed for rain manipulation studies in arid climates. An experiment identical to this one is installed in the Mojave Desert. Plant and animal community responses will be assessed in these studies.
Small mammal exclosures were constructed using 0.75 m tall welded wire fencing in a 90 × 90 m grid with four 30 m2 quadrants. The fencing is buried 25 cm below the soil surface and is 50 cm tall
A soil core study in a greenhouse and a companion field study is proposed to experimentally test the effects of precipitation timing and variation in soil traits on red brome and cheatgrass germination, growth and reproductive success. To examine these relationships at broader spatial scales under natural conditions, spatially explicit statistical models will be developed at the landscape scale. Landscape-level associations will be determined from feature datasets derived from remotely sensed and other geospatially explicit datasets within a Geographical Information System (GIS).
Using the paired transect experiment in the Beaver Dam Wash, we will monitor the effects of fire on reproductive success of nine native Mojave plant species in three pollination strategy categories: 1) wind pollinated, insect pollinated generalists) and insect pollinated specialists. Plant pollination and reproduction success will also be examined as a function of patch size and proximity to burn boundaries.
Each fall, 200 seeds of 15 plant species will be placed in a 25 cm2 plot within each of the 20 small mammal fence quadrants (10 burned, 10 unburned) (Fig. 4). Seed germination, growth and survival will be assessed. In addition, potted seedlings of the same 15 species will be placed within the experimental quadrants (Fig. 4) to test for folivory effects. The effects of grazing in burned landscapes will be experimentally tested using a network of 30 m2 livestock exclosures