Detecting and characterising land transformation in Laos
Due to its low population density, Laos still has vast environmental resources. Market and trade liberalisation began in 1985 after 10 years of socialism; in the 1990s, regional trade intensified and national policies encouraged a shift from subsistence to market-oriented production. The ensuing land use change throughout the country has been widespread and rapid. The main dynamics related to agricultural land use are changes in the land use practice of shifting cultivation by smallholders, and an expansion of permanent agriculture including annual crops and perennial plantations.
This land use change has profound social and environmental impacts that require appropriate policies by the government. However, spatial information on the changes has been limited, making policy targeting difficult. To support informed decision-making, CDE researchers have developed approaches (described below) to map land transformation.
Detecting changes in shifting cultivation
We assessed land use change using a dense time series of low resolution satellite images. The analysis showed that shifting cultivation is still widespread and quite dynamic in northern Laos. While there is an overall reduction in the area used for shifting cultivation, in some regions it has only just emerged. Reasons for this may be population pressure on the allocated land, the need to assure food security, or its use as an encroachment strategy to increase holdings of agricultural land.
Providing a socio-economic perspective on the changes
We analysed spatial data on the dynamics of shifting cultivation using auxiliary data to provide a socio-economic perspective on shifting cultivation landscapes. Most of these areas, located in remote uplands, show high poverty rates and are occupied by ethnic minorities.
Potential to enhance forest carbon stocks?
We combined the spatial data on shifting cultivation landscapes with a carbon inventory and village-level poverty data to determine the potential benefits of REDD+, the United Nations Collaborative Programme on Reducing Emissions from Deforestation and Forest Degradation in Developing Countries. The potential for implementing REDD+ lies in areas with high poverty rates and high carbon stocks. We were able to show that in Laos, a large share of the poor population lives in areas with low carbon stocks. Thus, in these regions, REDD+ cannot be considered a core instrument for poverty alleviation.
Shifting cultivation at district level
We used high resolution satellite imagery from one point in time to delineate shifting cultivation landscapes at the district level. This required a detailed classification of fallow land cover classes, which can be problematic in mountainous regions. Terrain and shadows also affect the spectral reflectance, decreasing the classification accuracy. By performing an object-oriented classification including texture measures, we were able to reduce this problem and classify cultivated plots and different types of fallow vegetation resulting from the shifting cultivation land use practice.
Mapping land cover changes
With a decision-tree classification approach we provided land cover maps for Xayabouri Province, at a 16-day interval from 2001 to 2011. Training areas for the development of the decision tree were derived from yearly medium resolution satellite images. Using a dense time series of low resolution imagery, we then classified the land cover, which provided spatially and temporally continuous maps for the study area. By performing a classification comparison of the 252 land cover maps, we could relate the different land cover change trajectories to land use and transformation processes, revealing the areas with stable land cover, deforestation (including expansion of permanent agriculture), forest regrowth, and shifting cultivation.