Construction Of Remote Sensing Inversion Method Of Photosynthetic Effective Radiation Based On Himawari-8 Satellite

Photosynthetically active radiation refers to the solar radiation between 400-700 nm,which directly affects vegetation photosynthesis and has an important impact on the global carbon-water cycle.At present,the photosynthetically active radiation estimation methods for satellites mainly include parameter empirical models,look-up table method,random forest and artificial neural network method.These methods all need some auxiliary data,which increases the uncertainty of the accuracy of the estimated photosynthetically active radiation.Himawari-8 is a new generation of geostationary satellites with a time resolution of 10 minutes,and can provide data for obtaining photosynthetically active radiation with high temporal resolution.Therefore,this paper proposes an optimization algorithm based on Himawari-8 satellite data,and only inverts the photosynthetically active radiation based on the apparent reflectance data of Himawari-8.Firstly,we perform sensitivity analysis use the E-FAST method and simulation accuracy validation of the SBDART radiative transmission model,and then use the optimization method to create a cost function for PAR inversion.The accuracy of the algorithm proposed in this paper is validated based on the data from two ground sites in Fukang and Beishawo,Xinjiang.The accuracy is also compared with the Himawari-8 PAR data.The accuracy of the optimized algorithm is analyzed from the daily and monthly changes.Finally,the photosynthetically active radiation in China was calculated based on the optimization algorithm,and the spatial distribution of photosynthetically active radiation of four months(January,April,July and November)in China was analyzed.Here are the main conclusions of this article:(1)The sensitivity analysis of the SBDART radiative transmission model by E-FAST sensitivity analysis method found that the SBDART radiative transmission model is most sensitive to the four parameters of ground surface reflectance,solar zenith angle,aerosol optical thickness and cloud optical thickness;(2)Verify the simulation accuracy of the SBDART radiation transmission model through 2016 measured data from the SGP and NSA sites of the ARM observation network,and find that the SBDART radiation transmission model can better simulate the downward shortwave radiation and its scattered radiation part and the upward shortwave radiation;(3)The daily and monthly changes of PAR retrieved by the created optimization algorithm are basically consistent with the measured PAR data at Fukang Station and Beishawo Station;in the verification of the accuracy of the daily variation,R at Fukang Station is greater than 0.9577,RMSE is lower than 51.86 W/?,R is greater than 0.9570 at Beishawo station,RMSE is 77.23 W/?.In the validation of monthly variation accuracy,R is larger than 0.926,RMSE is lower than 15.04 W/? at Fukang station.R is large than 0.894,RMSE is lower than 29.08 W/? at Beishawo station,respectively;(4)This paper analyzes that in the absence of clouds,the correlation coefficient between the PAR value calculated by the optimized look-up table algorithm and the PAR values observed at the two ground stations can achieve at 0.99 and 0.978,and the RMSE is 34.142 W/? and 28.03 W/?.Compared with Himawari-8 PAR product,the RMSE of 37.53 W/? and 79.07 W/? is greatly improved.