红外相机技术在鼠类密度估算中的应用

Estimating rodent density using infrared-triggered camera technology

2009年5月至2011年7月,在古田山国家自然保护区24 hm2(600 m×400 m)样地中用分层随机抽样法布置20台红外相机,监测样地内的鼠类密度。利用红外相机技术,引用物理学中气体分子碰撞率原理,在不对鼠类进行个体识别的情况下,估算样地内鼠类密度。结果表明,以此估算的样地内鼠类密度D3与标志重捕法估算的鼠类密度D之间不存在显著差异(P>0.05),两者契合程度高,说明此模型具有相当高的精确性。而在较高的相机分布密度(0.83台/hm2)下得出的鼠类密度季节消长状况也验证了该模型的可靠程度。

Conventional methods for investigation on densities of small mammals unavoidably produce environmental disturbance. By using infrared-triggered camera technology （ITCT）, animal densities could be obtained with minimal disturbance to animal＇s habitat. For large mammals, individually unique natural markings, such as spotted and striped fields can be used for recognition, but for small mammals, it is difficult to distinguish individuals, especially for nocturnal species. In this study, we used cameras triggered by infrared sensors to obtain images of passing animals to estimate the rodent density. Three biological indexes, that is, rodent＇s day range （or called speed of day movement） , distance and angle with which the camera detects animals are required for the modeling. Camera detection distance （0.01 km） and angle （50°） were provided by the infrared camera manufacturers in this study, whereas the rodent＇s day range had to be estimated in experimental sites. An experimental site of 600 m×400 m was selected at Gutianshan National Natural Reserve in Kaihua county, Zhejiang province. The site was divided into 18 areas based on the principle of stratified random sampling, and 20 infrared cameras were deployed in these areas. At the same time, the capture-mark-recapture method （CMR） was used to survey speed of rodent movement. The experiment was conducted from May 2009 to July 2011. The results derived from CMR, showed that the day range of rodent movement was approximately 100-300 m. By assuming the speed of day movement as v1 （ 0.1 km/day）, v2 （ 0.2 km/day）, v3 （ 0.3 km/day）, corresponding densities obtained from camera trapping were D1 , 02 and D3, respectively. When these three densities were compared with the density （D） derived from CMR, the results showed that D1 was highly significantly different （P〈0.01） from D. However, there were no significant differences between D2 and D （P = 0. 090） or between D3 and D （P = 0. 679）. The smallest difference was found between D3 and D. Therefore, D3 may represent the closest estimate. Taken D3 as the density estimate, rodentdensities was calculated for the period from May 2009 to July 2011 by using monthly independent photographs of rodents. Average density was （657± 81 ） individuals/km2. The seasonal changes in the rodent density which obtained from camera trapping with the high camera density of 83.3/kinz, confirm that our results are reliable. In this study, uninterrupted monitoring promoted efficiency of detection. Our results revealed that the density of rodent reached the peak from the beginning of May to the late summer because of abundant food resources and low interspecies competition. While from November to March, predation pressure and food shortage reduced rodent densities. As a consequence, intraspecific and interspecies competition was reduced, leading to a relatively stable community structure. In conclusions, our results demonstrate that ITCT with the aid of the capture-mark-recapture method can be used to reliably estimate rodent density with high accuracy and low costs.