Effect assessment of the project of grain for green in the karst region in Southwestern China:a case study of Bijie Prefecture

The karst region in Southwestern China possesses fragile eco-environment.With increasing population pressure and irrational land use,the local environment had experienced severe degradation showing as soil erosion,loss of soil fertility and deterioration of vegetation cover in the past decades.Since 2000,a series of ecological restoration projects,including the project of Grain for Green(conversion cropland into forest or pasture),have been implemented in this region.It is quite necessary and important to assess the effect of these projects so as to provide a scientific basis for decision-making of further ecological restoration projects.This paper takes Bijie Prefecture in Guizhou Province as a case study.Factors influencing the vegetation dynamics are identified as climate change and human activities.In order to verify the NDVI trends imposed by human influences such as the project of Grain for Green,this paper establishes a linear regression model between NDVI and climate by means of SPOT-VGT NDVI data and meteorological data from 1998 to 2001.By performing lots of linear regression calculations,the NDVI-climate influence model identifies accumulated precipitations from January to June and accumulated temperature from April to July as independent variables,and the NDVI from 200d to 210d per year which is assumed to be the max value of the year as the dependent variable.The model is also calculated at the pixels level with 1000m resolution.Then the residual trends of actual NDVI and simulated NDVI from 2002 to 2008 are calculated.Positive or negative trends expressed in the residuals are interpreted as human-induced restoration or degradation.Since vegetation condition is influenced by a series of ecological restoration projects in Bijie Prefecture,the effort of the project of Grain for Green should be distinguished from other ecological projects before evaluation.Combined with the result of residual analysis,the land-use data is also used.In particular,the paper evaluates the residual trends of cropland pixels which are extracted from land-use data because there is a strong correspondence between the project of Grain for Green and the recovery of vegetation on converted cropland.Finally the paper cites statistical data to validate the result.The results indicate that local vegetation condition has been significantly improved by the ecological projects since 2002.However,regional differences still exit in Bijie Prefecture.The trend of degradation has not been reversed in the east part such as Dafang county and Jinsha county.This suggests that the further ecological restoration projects are still needed in the karst region.The results also show that the converted land types of most restoration pixels are shrub land,grassland and cropland.91.68% recovery cropland pixels are located at the hills where cropland must be converted to woods or grass.Compared the number of recovery cropland pixels and the area of the project of Grain for Green implemented from 2002 to 2008 in each county,further analysis shows that there is an obvious positive relationship between the number of pixels and the area converted in the project of Grain for Green,except Hezhang county and Dafang county,which proved the effectiveness of the method.