Above-ground Biomass Model Estimation Based On MODIS And Meteorological Data In Xinjiang Grassland

Xinjiang is the third largest in China in terms of grassland’s area.Xinjiang grassland was pushed into an extremely unstable situation due to complicated landform and various climate types.MODIS is a good resource for biomass yield in large area due to its multi-temporal,multi-spectual and acquired frequently.In this thesis,the values of 7 kinds of Vegetation Index of Xinjiang grassland were extracted by ArcGIS,from 2010 to 2014.And above-ground biomass(AGB)estimation models of ten types grassland based on MODIS products were created by the statistical analysis in a soft called SPSS.And then,set up a model with average temperate(TEM)and total rainfall(PRE)from 2010 to 2014.At last,meteorological factors and remote sensing information were analyzed with AGB together and set up compound models named Remote sensing-meteorological-biomass model.The result showed that:1)The 7 kinds of vegetation index are change scientifically because different distributed of grassland.The total change trend is meadows are largest,the second is steppe,the third is desert,and the Northern Xinjiang larger than Sothern.2)There are significant relationship between AGB and 7 kinds of vegetation index of the 10 types grassland,remote sensing-biomass model based on them have good predictive effect.However,PRE and AGB in alpine meadow type were not significantly correlated.Same as TEM and AGB in temperate desert,.The rest of the grassland type’s AGB have significant relationship with TEM and PRE3)In this thesis,we study AGB model of azonal lowland meadow type,mountain meadow type,alpine meadow type,temperate meadow-steppe type,alpine steppe type,temperate steppe type,temperate desert steppe,temperate steppe-desert type,temperate desert type and alpine desert type in Xinjiang,the determination coefficient(R~2)of remote sensing and AGB model,meteorological and AGB model and remote sensing-meteorology and AGB model are bigger in alpine desert type,but the models precision is low because the small simple case.The three models of other types’R~2 lower than alpine desert type,but the precision are well.4)Due to complex topography and weather conditions volatile,AGB of 10 type’s grassland have different response to PRE and TEM.Alpine meadow type is related to TEM,and temperate desert type is related to PRE,when build AGB evaluation model of 10 type’s grassland use PRE and TEM,the accruey of alpine meadow type and temperate desert type’s model are lower.While build AGB evaluation model use the MODIS vegetable index based on rule of multiple linear regression model,we can received a model without the influence.5)The accuracy of models based on remote sensing variables were better than the models with meteorological variables for grassland,and the models with both of meteorological and remote sensing variables were the best from 2010 to 2014.Except the alpine desert type,q of remote sensing-meteorological and AGB model in other types were all great than 0.510.Thus,it is necessary to include both of remote sensing and meteorological variables while modeling.