We investigated the population dynamics of the endangered San Joaquin kit fox (Vulpes macrotis mutica) and factors that influence kit fox demographics at the Naval Petroleum Reserves in California (NPRC) during 1980-95. Kit fox abundance varied widely during this period and population estimates on a 216-km(2) study area ranged from 46 (0.21/km(2)) in 1991 to 363 (1.68/km(2)) in 1994. The finite rate of increase (lambda) ranged from 0.37 in 1995 to 2.22 in 1993. Of 341 adult and 184 juvenile foxes radiocollared, we recovered 225 and 142 dead, respectively. Mean annual survival rate for adults was 0.44 (range = 0.20-0.81) and the mean rate for juveniles from 1 May to 15 February (age 1) was 0.14 (range = <0.01-0.31). Coyotes (Canis latrans) and other predators were the primary source of mortality for adult and juvenile foxes, and vehicles were a secondary source. Mean annual reproductive success (proportion of females observed with pups in Apr-May) was 61% (range = 20-100%) for adults and 18% (range = 0-100%) for juveniles. Mean annual litter size was 3.8 and varied from 2.0 to 4.8. Population growth rates varied positively with reproductive success (P = 0.041), and both juvenile reproductive success and the proportion of new individuals were higher (P = 0.001 and P = 0.100) when the fox population was increasing. Leporids and kangaroo rats (Dipodomys spp.) were important foods for kit foxes, and annual use of foods by kit foxes varied markedly during the study Kit fox density was positively related to both current and previous year's composite prey indices (P = 0.001 and P = 0.067). Finite growth rates were positively related to rodent indices (P = 0.085) and current year's composite prey indices (P = 0.084). Rodent abundance and composite prey indices were strongly related to previous year's effective (Oct-Mar) precipitation (P = 0.002 and P = 0.003), as were kit fox density (P = 0.021) and <lambda> estimates (P = 0.002). A step-wise multiple regression model that included precipitation totals from the previous 3 years as independent variables explained 79% (P < 0.001) of the annual variation in fox abundance. Survival of adult kit foxes was negatively related to abundance of both coyotes (P = 0.046) and leporids (the primary prey of coyotes; P = 0.031). However, fox and coyote abundance generally tracked each other and neither kit fox abundance nor survival increased during a coyote control program conducted during 1985-90. We did not find strong evidence that oilfield activities were impacting kit foxes at NPRC. Fox abundance was significantly lower in oil-developed areas than in undeveloped areas during 6 of 13 years (P < 0.006). However, population trends of foxes appeared to be affected more by natural factors (e.g., food availability, competitor abundance) than oilfield activities. Kit fox population trends on NPRC appeared to be driven by precipitation-mediated prey abundance. Variation in prey availability, particularly kangaroo rats, produced significant and sometimes rapid changes in kit fox demographics and abundance. Kit fox abundance in the southern San Joaquin Valley is highly dynamic, warranting (1) conservation of sufficient habitat to maintain viable populations of foxes through population lows, and (2) research on management strategies to optimize habitat quality, particularly prey abundance.
