Assessing relationship between mammal occurrence and forest vegetation structure by using GWR

Forests play the most important role in the ecosystem, providing shelter, food sources and territories to floras and faunas. Forests are degraded from time to time, and forest conversion to agricultural plantations, especially for oil palm cultivation, has led to the problem of deforestation. Large-scale removal of trees changes species composition and wildlife occupancy as wild animals are sensitive to changes of landscapes. Preserving the primary forest is an excellent way to sustain the species richness. This study was conducted to gain in-depth insights into the existence of wildlife by assessing the numbers of species in two forests with different vegetation structure conditions: i) disturbed forest, and ii) undisturbed forest. A total of 120 camera traps were used to capture the images of the numbers of wildlife in both forest areas, from May 2015 to March 2016. To understand the driving factors that bring about wildlife existence within these two forest conditions, habitat quality attributes were measured: i) trees with diameter of above 45 cm at breast height (DBH) (DM45); ii) trees with DBH below 45 cm (DL45); iii) number of bamboo clumps (BMBO); iv) number of liana species (LANA); v) number of palm trees (PLMT); vi) number of fallen trees (FLTR); vii) number of saplings (SPLG); and viii) number of seedlings (SDLG). Two types of analyses have been conducted: i) general linear modelling; and ii) spatial analysis by using Geographical Information System (GIS). The aims of analysis with general linear modelling are to compare mammal species occurrence with the number of images captured; and establish the relationships between mammal species occurrence and the local factors in both the undisturbed and disturbed forests. On the other hand, the purpose of GIS analysis is to determine the explanatory variables that have non-stationarity effects on the mammal species occurrence. From the images captured, 3,730 small to large-size mammal species are identified. Of that, 15 of mammal species with small to large size were taken for analyses. From the general linear model, the undisturbed forest (n=2.683) has the highest mammal species occurrences compared with that of the disturbed forest (n=1.383). The images captured in the undisturbed forest are also high (n=50.87) compared with that of the disturbed forest (n=6.43). These comparative figures can be explained by the richness of biodiversity in the native forest. The mammal species occurrences are influenced by several factors; mammal species occurrence = f(number of lianas, trees with DBH >45 cm, number of palm trees, number of bamboo clumps, number of saplings). For GIS analysis, OLS Model 2 was examined and found to be the best model to determine the mammal species occurrences, based on the lowest AICc value of 210.81. The GWR Model 2 has been identified as the best method to determine the influence on mammal species occurrence (R2=41.56%), compared with GWR Model 1 (R2=32.87%), GWR Model 3 (R2=36.09%), OLS Model 1 (R2=25.86%), OLS Model 2 (R2=24.90%) and OLS Model 3 (R2=14.96%). The findings of this study about wildlife species richness and occupancy under different forest vegetation structures are indeed very valuable; the wildlife department can utilise this information as guidance for the purpose of conservation management and taking proactive measures.

Preview Text FH 2017 7 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item