| Author | Chomchid Imvitthaya |
| Call Number | AIT Diss. no.RS-11-02 |
| Subject(s) | Forest biomass--Remote sensing
|
| Note | A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Technical Science in Remote Sensing and Geographic Information Systems |
| Publisher | Asian Institute of Technology |
| Series Statement | Dissertation ; no. RS-11-02 |
| Abstract | In recent years, biogeochemical models have provided a robust way to assess the net exchange of carbon between the terrestrial biosphere and the atmosphere. The BIOME-BGC (Biome-BioGeochemical Cycles), one of the illustrious biogeochemical models used in forest studies, has been applied to find the NPP of different types of forest ecosystems. Nonetheless, one crucial limitation of the BIOME-BGC model is the accuracy of the ecophysiological input parameters for different forest types since determination of model parameter values in particular locations can be time consuming and difficult. One way to overcome this limitation, especially for forest which is difficult to access, is to combine remote sensing data and data assimilation in order to find the bestecophysiological parameters at regional scale. In this study, the BIOME-BGC model was calibrated to estimate the net primary production (NPP) of teak (Tectona grandis Lin F.), an important species in tropical deciduous forests, using data assimilation with remote sensing data (SPOT-LAI). A Genetic Algorithm (GA) was linked with BIOME-BGC (BIOME-BGC-GA) to determine the optimal ecophysiological model parameters. The sensitivity of ecophysiological parameters to LAI and NPP was examined, and 12 parameters which showed strong to medium sensitivity were selected for the optimization process. The sensitivity of GA parameters to optimize the result was also evaluated. The results of this investigation showed that crossover rate (0.7), with low mutation rate (0.09) and a large population size (100) gave the best optimization. The parameters optimized by BIOME-BGC-GA has significantly improved LAI simulation compared to the simulation by default parameters. The correlation (R2) between simulated LAI and SPOT-LAI increased from 0.57 to 0.78, and RMSE decreased from 1.40 m2/m2 to 0.53 m2/m2. The daily and annual-modeled NPP generated by optimized parameters and by default parameters were evaluated with daily MODIS NPP and field annual NPP, respectively. For the daily NPP, results showed that the modeled NPP using the optimized ecophysiological parameters can approximate the MODIS satellite daily NPP (R2=0.64, RMSE=1.11 gC/m2/d) better than the modeled NPP using default parameters (R2=0.19 and RMSE = 2.10 gC/m2/d). For the annual NPP, the optimized parameters improved difference between simulated and field from 16.31%-5.24% to 8.4-4.2%. It can be seen from the results that the annual modeled NPP using the optimized parameters is more stable. These improvements were mainly because the model’s optimized parameters reduced the bias of NPP by reducing the uncertainty of model parameters. The BIOME-BGC-GA therefore, can be effectively applied for teak forests in tropical areas. The study proposes an effective method of using GA to determine ecophysiological parameters at the site level and represents a first step towards the analysis of the carbon budget of teak plantations at the regional scale. |
| Year | 2011 |
| Corresponding Series Added Entry | Asian Institute of Technology. Dissertation ; no. RS-11-02 |
| Type | Dissertation |
| School | School of Engineering and Technology (SET) |
| Department | Department of Information and Communications Technologies (DICT) |
| Academic Program/FoS | Remote Sensing (RS) |
| Chairperson(s) | Honda, Kiyoshi |
| Examination Committee(s) | Surat Lertlum ;Voratas Kachitvichyanukul ;Nipon Tangtham |
| Scholarship Donor(s) | Royal Thai Government (RTG) |
| Degree | Thesis (Ph.D.) - Asian Institute of Technology, 2011 |