| Plateau pika (Ochotona curzoniae) is a keystone specie of the Tibetan plateau ecosystem, which mainly inhabit in the alpine meadow or the desert steppe at an elevation above 3000 meters. Degraded grassland is an ideal habitat of plateau pika, where the low vegetation condition provides an open horizon of the pikas. For providing refuges and food resource for many plateau species, plateau pika is very important for the stability of alpine meadow ecosystem, however they have been deemed to a pest on Tibetan plateau, as competing with the livestock for their food. Therefore, they had been killed with poison in some regions. For effectively control of pika's population, the scientists have carried out a series of studies since the 1970s, including morphology, behavior, ecology, physiology, biochemistry, molecular biology and so on. With these biological information about this specie, it can be helpful for us to build the spatial distribution model for their microhabitat-utilization in some area.As a central place forager plateau pika forages in a fixed location (e.g. dens and other refuges), researches have proved the habitat factors and the number of pikas' burrow entrances distribution are closely relateed. But how does this relationship worked? landscape scale studies on the distribution of plateau pika habitat utilization can be helpful to solve this problem. Consequently, we used normalized difference vegetation index (NDVI) and the plateau pika burrow entrances data to evaluate the vegetation utilization pattern of the pika in a 72.2km2 study area in Yunbo Gou, Shiqu County, Sichuan Province. The grazing forbidden policy has been carried out since 2007 in this area and the livestock competition has been removed since then. We carried out seven random line transects with a total length of 19km,from August to Suptember,2008. The NDVI of the study area was extracted from a CBERS-02 remote sensing image taken in September 18th,2008. The Slope imformation was extracted from the digital elevation model (DEM) based on a digitized topographic map of Shiqu County, then the NDVI and the slope were compared with the plateau pika burrow number in each sampling unit. A linear regression model showed that there is no significant linear relationship between the slope and pika burrow entrances (R2=0.04, P=0.441). In contrast, the value of the pika burrow entrances observed was significantly lower than the expected value in the low NDVI group(0.1-0.3), not significantly different in the middle NDVI group(0.3-0.4), and significantly higher than the expected value in the high NDVI group(0.4-0.6). A linear regression model showed an obvious positive correlation between NDVI and the pika burrow entrances y=16.50 x+0.87 (R2=0.78, P<0.001). The result shows that there is close relationship between the vegetation distribution and the number of pika burrow entrances, but no relationship between the number of pika burrow entrances and the slop in our study area.There are two types of burrow systems, one with complex structure, while another one with simple structure. After many years of digging, complex systems have many exits with multiple functions. Simple systems have less exit as a temporary shelters. For analyzing the special distribution pattern of different burrow systems, we introduced a new method to detect the complexity of small mammals'burrow structure. We injected the stage smoke into the burrows of pikas. The smoke fumed from connected burrows, and then we could find the quantity of exits and the location of every burrows. We analysed the complexity of burrow system by comparing the distribution of burrow entrance at the earth's surface. We studied the distribution of pika's burrow structure in different habitat (vegetation and slope) in a 72.2 km2 study area in Yunbo Gou, Shiqu County, Sichuan Province in September 2009. We classified the vegetation into 2 levels (higher and lower) and the slope into 3 levels (ground, gentle slope and steep slope) on the basis of previous research. In the different habitat we studied 144 pikas'burrow systems. The result show that the complexity of burrow systems are higher in high NDVI area than low NDVI area (Mann-Whitey U test, Z=-2.607, P=0.009). The complexity of burrow systems are higher in gentle slope than both ground (Mann-Whitey U test, Z=-4.116, P<0.001) and steep slope (Mann-Whitey U test, Z=-4.968, P<0.001),but there is no significant difference between ground and steep slope (Mann-Whitey U test, Z=-1.441, P=0.15). Linear regression model of slope and complexity of burrow systems is y=-0.02x2+0.39x+3.38, R2=0.534. The temperature in different habitat indicates that (℃):control (14.76±7.25)> ground (11.92±1.34)> gentle slope (10.89±1.69)> steep slope(9.57±3.63). The humidity in different habitat indicates that (%):ground (99.86±0.63)> gentle slope (97.75±3.94)> steep slope (93.51±8.78)>control (57.60±15.74). The result of Mann-Whitey U test shows that there are significant differences in all groups. The study shows that the complexity of pikas' burrow structure are different in different habitats. The burrows structure are more complex in the place where the vegetation condition is better. The distinction of the burrow structure in different slopes may be result in the changes of temperature and humidity. We found that more complex burrows are experienced in gentle slope. The main reason may be that the temperature and humidity here are more appropriate for pikas to overwinter and breeding.We built a model to reflect the distribution of pikas. We estimated that there are 416542 pikas in the study area of 72.2km. The density is about 57.69±10.50 individuals per ha. Pika population's density is 57.89±16.18 (n=5) 95% confidence interval basis on survey using Hayne transect method. The result extracted form model by arcGIS shows that there are 59.90±6.63 (n=50) individual per ha. The study indicated that the model prediction consistent with the actual. We tried to build a spatial distribution model of small mammal using 3S technique. This model will be helpful for the studies of pikas' behavior, home range, gene flow and so on. In the future, this model will have broad application prospects.
|
PDF Full Text Request