| Vegetation is a natural link to soil,atmosphere and moisture.To a certain extent,vegetation change can act as a "indicator" in global change research,that is,vegetation cover changes can reflect the trend of climate change from one side.By using the USGA LANDSAT5 data released from 2005-2011,the five vegetation indices of DVI,RVI,EVI,NDVI and TVI were extracted based on RS and GIS technology,and the corresponding hydrological data(precipitation,runoff depth)were extracted.Correlation and gray correlation analysis to study the relationship between vegetation cover change and water resources in the past 7 years in Guizhou Province.The main results are as follows:1.The precipitation of Karst River Basin in Gui Zhou province and landsat5 remote sensing image vegetation index(DVI,RVI,EVI,NDVI,TVI),through correlation analysis and greyrelational analysis method to explore the relationship.Generally speaking,There was a good correlation between vegetation index and precipitation in the study area.The correlation between average precipitation and DVI,RVI,NDVI,TVI was better than that in the same period.The correlation between precipitation and DVI,RVI,EVI,NDVI,TVI was higher than 0.8,and the index of the five vegetation was closely related to precipitation.In addition,by comparing the relationship between vegetation index and precipitation in March and in the year of,it is concluded that the correlation between the two is better than in the year of September in the year of March.2.There was a positive correlation between runoff depth and vegetation index,The correlation coefficients of runoff depth with DVI,RVI,EVI,NDVI and TVI in March were 0.682,0.770,0.717,0.662,and 0.637,respectively.The correlation coefficient RVI was the highest in September,and the correlation coefficient was 0.539;TVI and NDVI were the second,the correlation coefficients were 0.518 and 0.514 respectively.The correlation coefficient was 0.486 again,and the correlation of EVI was the most Poor,the coefficient is 0.372.Through the correlation analysis,the five vegetation indices of DVI,RVI,EVI,NDVI and TVI are closely related to runoff depth.3.With respect to the monitoring of precipitation,The cubic polynomial equation and the multivariate linear model in the curve model are more accurate than the quadratic polynomial.Compared with the quadratic polynomial,the curve model and the multiple linear regression model can reflect the relationship between vegetation index and precipitation.In the mathematic model of runoff depth and vegetation index,the exponential model is better than the multiple linear model,and the exponential model is higher than 0.8 in March and September.4.In the regression model of vegetation index and precipitation,the cubic polynomial model is superior to other curve models(quadratic polynomial and exponential regression model).Based on the RVI establishment of the cubic polynomial relationship model R2,March and September respectively reached 0.653,0.774,and the third polynomial model validation,found that the measured value and the forecast is very close to the average error of 18% and 7%.The R2 of the cubic polynomial relation model based on RVI is 0.774,and the cubic polynomial model is verified.It is found that the measured value is very close to the predicted value,and the average error is 7%.In summary,it is shown that the cubic polynomial model can meet the needs of macroscopic monitoring of water resources in karst areas,and it is also a simple and effective method to monitor the precipitation by the cubic polynomial relationship model.In the mathematical model of vegetation index and runoff depth,the multivariate linear model and exponential model R2 are all larger than 0.8.5.The difference between the vegetation index and the water resources is different,and the vegetation index and the runoff depth are higher than the precipitation and vegetation index.Among them,the best correlation with runoff depth and precipitation is the ratio vegetation index(RVI). |
PDF Full Text Request