Characterizing Tropical Forest in Kinabalu Park, Sabah, Malaysia, Using Landsat-TM Data
Analysis and Discussion
The spectral response (DN) were examined in a tasseled cap feature space. Plane of vegetation is defined by the features of brightness and greenness, as x-axis and y-axis, respectively (figure 1). In this plane, trajectories in response to the change in the percentage crown cover wore observed. With high brightness and low greenness, bare soil is located in the lower right part of the feature space. With increasing percentage crown cover, the vegetation pixels traverse from the lower right part of the plane to the group of pixels with closed in the upper left part of the plane.
Figure 1. Plane of Vegetation
In the upper left part of the feature space, influence of the crown diameter on the spectral response from closed forest (
³ 75% crown cover ) was observed. When the crown diameter is small, sub alpine forest tends to occupy the part that nearest to the origin. In contrast, big crown forests ( e.g. lower montane ) tend to distribute farther away from the origin.
The actual relationships between the forest variables and the Landsat-TM data were investigated by regression analyses. The percentage crown cover is strongly related to all the remotely sensed data except band 4 (Table 1). Only a weak positive relationship (R
2=0.27) is found between the percentage crown cover and the near infrared data. This is attributed to the influence of the bare granite and variations in the vegetation amount associated with the different types of vegetation.
| Forest variables | Band 3 | Band 4 | Band 5 | Band 7 | Brightness | Greenness | Wetness |
| Percentage crown cover | 0.847 | 0.27 | 0.951 | 0.984 | 0.809 | 0.869 | 0.984 |
| Crown Diameter | 0.607 | 0.718 | 0.587 | 0.407 | 0.63 | 0.535 | 0.691 |
Table 1. regression Coefficients between forest variables and the Remotely sensed data
In contrast, the relationships between the percentage crown cover and the near infrared bands(bands 5 and 7), and also between the forest variable andwetness are vary strong (Figure 2). This is because the increases in the percentage crown cover have led to the vegetation moisture content and the soil moisture that absorb solar radiation in bands 5 and 7. wetness is sensitive to the soil moisture and canopy moisture content (Lillesand and kiefer, 1994). The increases in the soil and vegetation moisture content due to increasing vegetation cover are clearly shown by the very strong positive curvilinear relationship between the percentage crown cover and wetness. A strong negative relationship is also found between the forest variable and band 3(R
2=0.847). band 3 is designed to sense in the strong absorption region (red ) of chlorophyll and b. the increases in the percentage crown cover also mean increases in the green vegetation amount and thsis the amount of the chlorophyll.
Figure 2. relationship between Percentage Crown Cover and the Landsat-TM Data (bands 5 and 7 and wetness)
