Monitoring And Characteristics Of Swidden Agriculture Using Sentinel-2 In Northern Laos

Swidden agriculture is an ancient,widespread and controversial form of traditional self-sufficiency agriculture in tropical mountainous areas.It is a complex and dynamic agroforestry system.Since 2008,Swidden agriculture has become an important issue of the UN’s "Reduce Greenhouse Gas Emissions from Deforestation and Forest Degradation"(REDD / 2008)program,and its development status is closely related to multiple UN Sustainable Development Goals(SDGs).The problems of carbon emission and carbon cycle caused by swidden agriculture have become the focus of academia and are highly valued at the policy level.Aiming at the lack of remote sensing research on the exact distribution,scale and spatio-temporal characteristics of swidden agriculture,the existing research is mainly limited to local scale,and the high-precision monitoring of swidden agriculture is still to be strengthened.This article first quantitatively evaluates the availability of Sentinel-2 A / B images in swidden agriculture monitoring.Based on the satellite images of high spatial resolution of 7159 scenes available from Sentinel-2 A / B in 2016-2019,a set of collections has been perfected Phenological characteristics,spectral attributes,multi-step threshold method,NDre1 m index algorithm for remote sensing monitoring of swidden agriculture.Then,the typical tropical mountainous area of swidden agriculture in northern Laos is taken as the research object to discuss the relevant characteristics of swidden agriculture.Based on this,the following conclusions are drawn:1)It is determined that the appropriate threshold for Sentinel-2 A / B image cloud coverage is 20%.Under this threshold,the Sentinel-2 A / B historical image cloud in northern Laos has a high probability of obtaining images in the dry season,a low probability of obtaining images in the rainy season,and a western province in the dry season The probability is much higher than that in the east,and the probability of acquiring images in the western part of the rainy season is slightly lower than that in the eastern region.Quantitative evaluation of the compatibility and continuity of Sentinel-2 A / B images in remote sensing monitoring,and the higher spatial and temporal resolution of Sentinel-2 A / B images and the high probability of acquisition of dry season images provide advantages for remote sensing monitoring of swidden agriculture Data foundation.2)Among the 22 spectral indexes,NDVIre2,NDVIre2 n,NDVIre3,and NDre3 n can not identify the burning land and natural forest.In addition to the 17 spectral indexes,indexes with a spectral interval greater than 1,the ability to identify fires from high to low is SR,SRre2,SRre1,CIre,MSRre,and MSRren.The index of the spectral interval in the order of less than or equal to 1 from high to low is NBR,NDre1 m,NDVI,NDre2 m,NDre2,NDre1,NDVIre1,NDVIre1 n,GNDVI,MNDre,DVI.Among the pure red side spectral index,NDre2 recognizes the highest sensitivity,part NDre1 m has the best recognition effect in the red edge index,and the strongest non-red edge index is NBR.Combined with the multi-step threshold and accuracy verification of the 22 spectral indexes,the NDre1 m index is the optimal index for identifying the characteristics of swidden agriculture.3)The area of the swidden agricultural area in northern Laos is 8426.34 km2,accounting for 5.45% of the total area in northern Laos and 78.48% of the cultivated area in northern Laos.The province with the largest swidden farming area is Xayaboury,which accounts for 16.81% of the swidden farming area in northern Laos,followed by Luangprabang,Vientiane,Polikansai,Vientiane Capital,Huaphanh,Oudomxay,Champasack,Phongsaly,Xaysomboon,Bokeo and Luangnamtha.The area of swidden agriculture in the central region is slightly larger than that in the eastern and western regions.Swidden farming occurs at altitudes below 2300 m,with the largest area around 170 m.The slope appears at 0 ~ 65 °,and the peak appears at about 5 °.This article provides a certain reference for the research and development of Sentinel-2 data in the field of swidden,and provides a prerequisite data foundation for more accurate swidden agriculture monitoring algorithms and spatial distribution information for the research on the impact of the dramatic changes in swidden agriculture.