The forest-swidden-plantation (FSP) nexus in the tropics
First, there is a need to carry out a synergic analysis of the tropical forest-swidden-plantation (FSP) nexus more than at any time in history under the context of rapid deforestation. The FSP nexus integrates both natural/secondary forest and commercial plantation as the anterior and posterior ends with swidden agriculture (also swiddening practice) as the main thread. This integration contributes to understand the self-evolution of swidden agriculture (including fallow period and intensity), and also is helpful to investigate its corresponding roles in natural forest changes (gain or loss) and expansion of tropical plantation(Ziegler et al., 2009), such as tree crops (e.g. rubber, coffee, and pulp trees) and non-tree crops (e.g. banana, pineapple and sugarcane). Although recent studies have enriched significantly our knowledge of forest loss(Curtis et al., 2018; Hansen et al., 2013) and plantation expansion(Peltzer, Bellingham, Dickie, & Hulme, 2015; Ziegler et al., 2009), the previous research has proceeded separately and yet to be integrated under the umbrella of sustainable development of swidden agriculture. In particular, much less is known about the processes, mechanisms, and future trends of the interplays of natural and secondary forests changes, swidden agriculture evolution and transformation, and commercialized plantations expansion in the tropics (see the figure, panel B-I ).
Second, there is a need to integrate remote sensing methods with more traditional research. Consistent mapping of annual dynamics of swidden agriculture (including fallow cycles) and cash crops plantation facilitates tracking the advance trajectories of agricultural expansion or forest loss (retreat) in TAFF. With the free access of satellite imagery including Landsat and Sentinel-2, more efforts and new algorithms are necessary to map and investigate the dynamics of swidden-fallow trajectories, forest gain and/or loss, and plantation expansion. People often question the data availability of optical satellite imagery (e.g. Landsat) for mapping swidden agriculture and rubber plantations in the tropics especially during the growth periods, usually overlapping with the rainy season. However, this situation is totally improved within a half-year dry season due to the tropical monsoon climate(Li, Feng, & Xiao, 2018). Notably, the rotational and commonly-used slash-and-burn phases (especially the signal of fire) provide a key window for monitoring the transition from primary forest to monoculture using phenology-based methods and/or machine learning algorithms (e.g. random forest and neural network).
Finally, there is a need to understand the mechanisms through phenomena of forest change (esp. deforestation), swidden transformation, fallow period shortening, and plantation advance. Usually, the variation in drivers and/or mechanisms of the FSP dynamics are huge in different parts of the tropics. As for drivers of global deforestation, for instance, there are agro-industrial crops, plantations (including oil palm), pasture, small-scale clearing, selective logging, fire, and infrastructure/natural disturbance(Seymour & Harris, 2019). Population growth and migration, livelihoods diversity, markets and infrastructure development, public policies and attitudes and other socio-cultural drivers influence the persistence or demise of swidden agriculture(van Vliet et al., 2012). Offsetting carbon, increasing bioeconomy demand, urban migration, and government programs serve as the potential drivers of tropical plantation expansion(Rudel, 2009). Geo-economic cooperation mechanisms, national policies on land allocation, forestry, agriculture and natural resources, and local (including household level) agricultural practices are comprehensively needed to probe into the mechanisms through a bottom-up inductive approach.