Development of canopy properties using lidar and high resolution aerial photograph for simulating throughfall in forested area

Abstract

Mapping high-resolution throughfall in the tropical forest by ground measurement is impractical due to several prominent factors including thick and dense canopies which give rise to several constraints such as physical access, induce threat from animals, and requiring a long sampling period for alarger area. As such, using LIDAR and high-resolution images is an alternative that provides the ability to measure canopy properties at a fine scale. Nonetheless, an appropriate and operational approach is still lacking. Many throughfalls or interception studies utilizing remote sensing data were either focusing on a regional scale (>500 m), two-dimensional perspectives (2D images), or limited to the assessment of the vertical canopy thickness or properties only. Therefore, this research was aimed to initiate a simplified and practical method to estimate the throughfall by the function of the canopy properties in the tropical forest of Limbang, Sarawak. This study utilized the volumetric canopy density (VCD) to represent the canopy properties that characterize the interception process. The VCD was derived by two major inputs, (1) the horizontal canopy closure (HCC), and (2) volumetric canopy depth (VCT) from LIDAR and high- resolution images respectively. The VCD was used to modify the canopy storage component in the Gash interception analytical model. The throughfall estimates were obtained from the modified equation and later calibrated using the localized Dykes throughfall model applied in nearby sites. The derived HCC and VCT showed from good to a moderate agreement with the in-situ measurement at a correlation of 0.638 and 0.522 respectively. The correlation between LIDAR derived throughfall and the simulated throughfall based on in-situ biophysical data was about 0.765. The average quantitative error was 0.01 mm/hr. Validation against the in-situ throughfall in the nearby sites of Batu Apoi Forest reserve showed the LIDAR derived throughfall produced a good correlation (r² = 0.9703) and a low error (1.2 mm/hr). This study has demonstrated the capability of LIDAR and high-resolution images to provide an effective mapping of the high-resolution throughfall.