The Examination of The Satellite Image-Based Growth Curve Model Within Mangrove Forest

Abstract

Developing growth curve for forest and environmental management is a crucial activity in forestry planning. This paper describes a proposed technique for developing a growth curve based on the SPOT 6 satellite imageries. The most critical step in developing a model is on pre-processing the images, particularly during performing the radiometric correction such as reducing the thin cloud. The pre-processing includes geometric correction, radiometric correction with image regression, and index calculation, while the processing technique include training area selection, growth curve development, and selection. The study found that the image regression offered good correction to the haze-distorted digital number.  The corrected digital number was successfully implemented to evaluate the most accurate growth-curve for predicting mangrove.  Of the four growth curve models, i.e., Standard classical, Richards, Gompertz, and Weibull models, it was found that the Richards is the most accurate model in predicting the mean annual increment and current annual increment.  The study concluded that the growth curve model developed using high-resolution satellite image provides comparable accuracy compared to the terrestrial method.  The model derived using remote sensing has about 9.16% standard of error, better than those from terrestrial data with 15.45% standard of error.