Portal de Periódicos da CAPES

Correlation Analysis between NDVI and Climatic Factors of Grassland Ecosystems in the Northern Tibetan Plateau from 1982 to 2003

2009; Chinese Academy of Sciences Geography; Linguagem: Inglês

1007-7588

Zhenxi Shen,

Environmental Changes in China

Normalized difference vegetation index (NDVI), which is closely related to plant absorbability of photosynthetically active radiation (PAR), is generally taken to be a good indicator of terrestrial vegetation. We used time series (1982-2003) NDVI from the Global Inventory Modeling and Mapping Studies (GIMMS) in combination with climatic data from eight meteorological stations to examine the relationships between climatic factors and NDVI for grasslands of different types across the Northern Tibetan Plateau. Also, we discussed contributors affecting the correlation between NDVI and climatic factors. A bivariate correlation analysis was performed to examine the linear correlations and lag-linear correlations between monthly maximized NDVI and 6 monthly climatic factors, including mean monthly air temperature, mean minimum air temperature, mean maximum air temperature, relative humidity, mean wind velocity and cumulative precipitation. The partial correlations between temperature/precipitation, temperature/precipitation/wind velocity and NDVI were investigated by a partial correlation analysis in order to remove errors resulting from interactions among these climatic factors. Except the Gerze and Gaer stations, there were significant positive linear correlations between NDVI and temperature, precipitation and air moisture of the current month (P0.001) in other six stations. The climatic factor having the highest correlation with NDVI was mean monthly minimum temperature. Correlation coefficients between NDVI and mean monthly minimum air temperature for the Suoxian, Nagqu, Amdo, Damxung, Baingion and Xainza stations decreased in sequence. Alpine meadow showed the highest positive correlation with monthly cumulative precipitation than did grasslands of other types. Mean monthly wind velocity was negatively correlated to NDVI (i.e. Damxung, P0.05;other stations, P0.001). NDVI of the Gerze station showed the highest correlation with wind velocity compared with other climatic factors of the current month. Regarding grasslands of all types across the Northern Tibetan Plateau, the response of NDVI to temperature and precipitation showed a time lag effect. Except the Gaer station, the time lag as to responses of NDVI to precipitation and temperature for grasslands of different types was around one month and one to two months or so, respectively. The accumulated time lag periods of NDVI and corresponding precipitation was two months for the Suoxian and DamXung stations, four months for the Gerze and Gaer stations, and one month for the other four stations. The results of partial correlation analysis showed that temperature was the most important factor affecting NDVI of grasslands, and wind was another key factor affecting NDVI of grasslands. In general, changes in NDVI for the Suoxian, Nagqu, Baingion and Xainza stations were primarily driven by temperature-wind velocity-precipitation, but for the Amdo and Damxung stations, that relation was changed to temperature-precipitation-wind velocity. It was concluded that vegetation type, altitude, mean annual wind velocity and mean annual precipitation have significant influences on the correlation between NDVI and climatic factors.