Changes in Philippine Vegetation and Associated Impacts of Temperature and Precipitation

The sequestration of carbon dioxide in vegetation is recognized as a key mitigation strategy for overcoming the impact of greenhouse warming. The Philippines is estimated to have 16 M ha of 'forest land', which consists mainly of natural primary forest, a large fraction of which has been degraded to grassland areas. To evaluate the effectiveness of Philippine vegetation as a sink for carbon dioxide it is important to have an inventory of land cover types including existing forested areas. However, much of available data are field measurements with incomplete historical records and sparse spatial coverage. To overcome this limitation, we analyzed Normalized Difference Vegetation Indexes (NDVI) satellite data from the Moderate Resolution Imaging Spectroradiometer (MODIS) to establish the state of the vegetation in the country and determine how it has been changing during the last decade. Using 16-day composites of NDVI at 250 m spatial resolution, we create time-series plots over the course of eleven years. Large seasonal and interannual variability were observed reflecting the differences in the type of vegetation in different parts of the country. NDVI maxima are observed during the months August to October with moderate interannual variability while NDVI minima are recorded in the summer months (May to June) with much stronger interannual variability. The lowest value of NDVI from 2000 to the present was observed in the summer of 2010, with a decline of 5% from the previous year compared to an average fluctuation of 2%. This event coincided with an anomalously warm and dry year in 2010. Using temperature data from the same satellite, we find a correlation coefficient of 0.78 between temperature and NDVI in 2010 compared to the average correlation coefficient of 0.33 over the entire observation period. In addition, using rainfall data from the Tropical Rainfall Measuring Mission (TRMM) we get the highest correlation coefficient for the same year between precipitation and NDVI with a value of 0.80 compared to a mean of 0.58. These strong correlations of NDVI with both temperature and precipitation are good indications that these physical variables drive the 2010 anomaly in vegetation. Further analyses of the data also show that the correlations of NDVI with temperature and precipitation are generally higher during the summer season than the rest of the year, with maximum correlation coefficients of 0.74 and 0.65 for temperature and precipitation, respectively. This trend suggests that the vegetation cover in the Philippines is more sensitive to the warm and dry weather that characterizes the summer season - a situation that could be exacerbated by global warming. The vegetation time-series plots also exhibit a noticeable degree of variability in the length of growing and drying periods, which ranges from 111 to 191 days of growing period and 141 to 285 days of drying period. These periods correspond to the number of days between the recorded maximum and minimum NDVIs for any given year. Overall, it is apparent that an extended drying period is accompanied by a delayed onset of rainy season and shorter growing season.