| China is the world’s largest rice producer,and rice output occupies an important position in food security system and agricultural production.In recent years,affected by the regulation of market mechanisms and the acceleration of urban processes,the rice planting area and planting systems in different regions have changed,which in turn affects food security.In this study,we selected the dominant rice planting area in southern China—Poyang Lake area as the study area.Based on data from multi-source,multi-temporal remote sensing images,field observations,and socio-economic statistics,with the ESTARFM constructs a high-temporal-spatial-resolution data,and explores methods and rules of extracting paddy multiple according to the biological and spectral characteristics analyze the temporal and spatial pattern of rice ripening system,and clarify the driving factors of its change.The conclusions are as follows:(1)The ESTARFM algorithm has good applicability to areas with fragmented land and strong landscape heterogeneity,and has important practical significance for resource and environmental monitoring in southern,China.The sensitivity of the parameters shows similar characteristics in the band and vegetation index fusion.As the ,m and N increase,the fusion error shows a trend of first decreasing and then becoming stable or increasing.When using the spatio-temporal fusion method in different regions,parameter sensitivity analysis is required.Based on Landsat and MODIS multi-temporal images,the ESTARFM algorithm is used to generate hightemporal-resolution images to replace the cloud and cloud shadow areas in the Landsat image.It can effectively remove the "cloud pollution" phenomenon of images.(2)Based on the ESTARFM algorithm to fuse Landsat and MODIS images to obtain a high-temporal-spatial-resolution data set,use the spectral and phenological characteristics to identify single-cropping rice and double-cropping rice.The user accuracy is higher than 86%,the mapping accuracy is higher than 83%,and the overall accuracy and Kappa coefficient are 91.24% and 0.88,respectively.In addition,compared with the WESTDC,the results are rich in spatial information,which can capture more detailed fragmented rice field area information.These results show that the FLMP algorithm can effectively obtain high-precision paddy multiple information,and it has good applicability in the complex landscapes of southern.(3)The characteristics of the temporal and spatial changes of different paddy multiple in Poyang Lake area from 2004 to 2019: The paddy planting area has shown a continuous downward trend,and the planting mode has changed from single-cropping rice in 2004 to single-and double-cropping rice in 2019.In terms of spatial distribution,the reduced area of single-paddy planting was mainly located in the Fuhe River Basin and Xinjiang River Basin,and the increased areas were mainly located in the Xiushui River Basin and the northern Poyang Lake area.The overall area was mainly reduced;the double-paddy planting area reduced area in the Xiushui Basin and the northern Poyang Lake area,and the increasing area in the Ganjiang River Basin and the southern Poyang Lake area.The overall area is mainly increased.During the study period,singleand double-cropping rice changed frequently,the change was mainly from singlecropping rice to double-cropping rice("single to double"),located in traditional singleand double-cropping rice cropping areas in the southern Poyang Lake area and Fuhe River Basin.The conversion of double-cropping rice to single-cropping rice("doubleto-single")is mainly distributed in the Xiushui Basin and the northern Poyang Lake area.(4)Based on the random forest model and geographic detectors,the relative influences and interactions of driving factors of rice maturation changes in the study area were discussed: Rice yield,soil type,total power of agricultural machinery such as per capita agricultural income level have a greater impact on the temporal and spatial changes of rice ripening system.In contrast,driving factors such as rural labor resources,rural population,and related agricultural policies have less influence on the changes in rice ripening system;Pairwise driving factors Most of the interactions formed between the two factors and non-linearly enhanced the explanatory power of the driving force of rice maturation,and the greater the degree of influence of the driving factors,the greater the value of the interaction,which further indicates that the rice in the study area The maturation system is formed by a combination of multiple driving factors.Among them,land quality,agricultural infrastructure and technical level,and per capita agricultural income are the main driving factors for the temporal and spatial changes of rice maturation;Based on several main factors affecting the changes in rice maturation in this area The driving factors(topography,field quality,water source issues,etc.)respectively put forward policy recommendations. |
PDF Full Text Request