NPP Estimation Using GF-1 Data In Semi Steppe Area

Net primary productivity?NPP?is an important ecological indicator to evaluate ecosystem,it plays an important role in the global carbon balance and evaluation,Working out effective research to get high precision NPP result become main task for native and national experts.Traditional manual measurement to estimate NPP is time-consuming and strenuosity.It is just suitable for small scale estimation except large area.With the development of remote sensing technology,it entered a new phase,many scholars began to explore the remote sensing monitoring methods to estimate NPP,which is mainly divided into physiological and ecological process model,climate productivity model,light energy efficiency model and the ecological coupling model.The light energy efficiency model is one of the common models in these methods,it has clear physical meaning.The input parameter is also simple and easy to obtain.It provide an adequate technical support for the NPP remote monitoring in arid and semiarid regions.The vegetation in Arid and semi arid areas is sparse,the surface conditions are relatively complex,grassland is the main land cover use.But the existing data is coarse resolution and influenced by soil background greatly,making it different to achieve high precision result.for this problem,GF-1 16 m data was used for NPP estimation in semi-arid area combined with real data.CASA model was used in this process.Key parameters were optimized in the model to make it fit to semi area and GF-1 data.Zhenglan Banner of Inner Mongolia was selected by study area and achieved remote sensing data in 2014.Study proposed that monthly time-series data was used to land cover classification.The maximum energy use efficiency???,maximum NDVI,minimum NDVI were obtained from estimating using the land cover classification combined with measured data and meteorological data in the same time.Then proportion of FPAR_NDVI and FPAR_SR was optimized.Finally,NPP result was eatimated using modified CASA model and it was analyzed and verified by measured data.for the purpose of ecological monitoring in Zhenlan Banner,providing the scientific basis for the use of GF-1 data.The principal research results are as follows:?1?GF-1 data have high spatial and high temporal resolution characteristics,it is useful to distinguish land cover types in semi-arid areas based on NDVI time series data,The results showed that the accuracy was 83.37% and Kappa coefficient was 0.79 using original spectral image combined with the NDVI time series data,it is improved by about 10% compared with only using original data.?2?Before using GF-1 data to estimate NPP,largest light energy utilization efficiency of the grassland in Zhenglan Banner was inverted based on measured data.The value of steppe was 0.518gC/MJ.Adjustment parameters was 0.378;largest light energy utilization efficiency of the humidiherbosa was 0.523gC/MJ,Adjustment parameters was 0.402;largest light energy utilization efficiency of the meadow was 0.52gC/MJ,Adjustment parameters was 0.387.and it was different between the ever value 0.5.then use the parameters achieved above to estimate NPP by GF-1 satellite data.R2 between the results and the measured data reached 0.71,proving that NPP estimation by GF-1 data in semi-arid area is feasible.?3?The result shows that NPP spatial distribution in Zhenglan Banner is consistent with the vegetation types and climate characteristics.Total annual NPP is 1.43 × 10¹² gC / a,the average NPP in July is 33.86 gC / m².Grassland is the main accumulation type for NPP,indicating that grassland resource is a positive abundant in the Zhenglan Banner.The contribution rate reached 63.76%.The month maximum NPP of Zhenglan Banner appeared in July among the hole 2014 year.NPP accumulation increased rapidly from the beginning of April and basically ended in October.