Modeling seasonal surface temperature variations in secondary tropical dry forests.
| Publicado en | International Journal of Applied Earth Observation and Geoinformation, v. 62:122-134 |
|---|---|
| Autores | Cao, S., Sanchez-Azofeifa, G.A. |
| Año de publicación | 2017 |
| DOI | https://doi.org/10.1016/j.jag.2017.06.008 |
| Afiliaciones | Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, T6G 2E3, Canada |
| Programa | CRN3 |
| Proyecto | CRN3025 |
| Keywords | |
Highlights
•Seasonal LST variations were characterized for a secondary TDF.
•Thirty-eight Landsat-8 TIRS data were used to model LST time series of 3 years.
•The modeled LSTs were related to 7 VIs in three successional stages.
•Precipitation and soil moisture data obtained from a WSN were analyzed.
•We found that precipitation ultimately determined the LST variations in TDFs.
Abstract
Secondary tropical dry forests (TDFs) provide important ecosystem services such as carbon sequestration, biodiversity conservation, and nutrient cycle regulation. However, their biogeophysical processes at the canopy-atmosphere interface remain unknown, limiting our understanding of how this endangered ecosystem influences, and responds to the ongoing global warming. To facilitate future development of conservation policies, this study characterized the seasonal land surface temperature (LST) behavior of three successional stages (early, intermediate, and late) of a TDF, at the Santa Rosa National Park (SRNP), Costa Rica. A total of 38 Landsat-8 Thermal Infrared Sensor (TIRS) data and the Surface Reflectance (SR) product were utilized to model LST time series from July 2013 to July 2016 using a radiative transfer equation (RTE) algorithm. We further related the LST time series to seven vegetation indices which reflect different properties of TDFs, and soil moisture data obtained from a Wireless Sensor Network (WSN). Results showed that the LST in the dry season was 15&ndash20 K higher than in the wet season at SRNP. We found that the early successional stages were about 6&ndash8 K warmer than the intermediate successional stages and were 9&ndash10 K warmer than the late successional stages in the middle of the dry season meanwhile, a minimum LST difference (0&ndash1 K) was observed at the end of the wet season. Leaf phenology and canopy architecture explained most LST variations in both dry and wet seasons. However, our analysis revealed that it is precipitation that ultimately determines the LST variations through both biogeochemical (leaf phenology) and biogeophysical processes (evapotranspiration) of the plants. Results of this study could help physiological modeling studies in secondary TDFs.