Relationships between tree-ring width index and NDVI of grassland in Delingha

Using five well-replicated Qilian juniper (Sabina przewalskii Kom.) tree-ring width index se- ries, monthly normalized difference vegetation index (NDVI) of grassland, and climatic data from 1982 to 2001, the relationships between tree-ring width index, NDVI of grassland, and climatic data were analyzed firstly. Then, the relationship between tree-ring width index and NDVI of grassland was explored. The re- sults showed that: (1) Temperature and precipitation in June influenced tree-ring width index and NDVI of grassland deeply in Delingha. (2) There were sig- nificant relationships between five tree-ring width index series (DLH1-DLH5) and monthly NDVI of grassland from June to September, with the most significant relationship being between tree-ring width index series and NDVI of grassland in August. (3) The PC1 (the first principal component derived from DLH1-DLH5 series) exhibited good agreement with monthly NDVI of grassland in the grass growth sea- son (from June to September) and the averaged NDVI in the growth season, which was attributed to their common responses to water-supply limit in Delingha. This study may allow an increase in studying the past dynamics of grassland in Delingha in that the variation of grassland NDVI during the last millennium has been reconstructed from PC1.

Using five well-replicated Qilian juniper （Sabina przewalskii Kom.） tree-ring width index series, monthly normalized difference vegetation index （NDVI） of grassland, and climatic data from 1982 to 2001, the relationships between tree-ring width index, NDVI of grassland, and climatic data were analyzed firstly. Then, the relationship between tree-ring width index and NDVl of grassland was explored. The results showed that： （1） Temperature and precipitation in June influenced tree-ring width index and NDVl of grassland deeply in Delingha. （2） There were significant relationships between five tree-ring width index series （DLH1-DLH5） and monthly NDVl of grassland from June to September, with the most significant relationship being between tree-ring width index series and NDVl of grassland in August. （3） The PC1 （the first principal component derived from DLH1-DLH5 series） exhibited good agreement with monthly NDVl of grassland in the grass growth sea- son （from June to September） and the averaged NDVl in the growth season, which was attributed to their common responses to water-supply limit in Delingha. This study may allow an increase in studying the past dynamics of grassland in Delingha in that the variation of grassland NDVl during the last millennium has been reconstructed from PC1.