The history and revival of swidden agriculture research in the tropics

Abstract

Swidden agriculture used to occur in the temperate zones and currently occurs almost exclusively in the tropics. Academic research on the millennium-long farming system did not occur until the mid-eighteenth century, followed by scattered and sporadic research work before the twentieth century. So far, a thematic review of the history of swidden agriculture research based on the Institute for Scientific Information (ISI) Web of Science, which provides the longest coverage from 1900 to the present, has not yet been reported. The lack of a 20th-century literature review restricts understanding the revival of swidden agriculture research. With the journal publications (including Articles, Review Articles, and Data Papers) indexed by Web of Science and Google Scholar, we divided the history of swidden agriculture research into three developmental stages: descriptive transcription, critical analysis, and comprehensive analysis, with the years of 1957 and 2008 as the watershed years, respectively. Notably, 2008 emerged as a watershed year for the revival of swidden agriculture research in the tropics. Launching and implementing the United Nations Collaborative Programme on Reducing Emissions from Deforestation and Forest Degradation in Developing Countries promotes top-down rethink and rediscovery. In contrast, the free Landsat archive provides bottom-up support for consistent historical satellite observations. The synchronic emergence of the UN-REDD Programme and the free Landsat data policy may be coincidental. Yet, their combination and the global economic crisis since 2008 have become a catalyst and impetus for putting the longitudinal and horizontal analyses of swidden agriculture together. After a century of debate, swidden agriculture is gaining the academic attention it deserves.

Introduction

Swidden agriculture, variously known as shifting cultivation or slash-and-burn farming (Conklin 1957; Kleinman et al. 1995), is an ages-old agricultural technique (Ickowitz 2006) and an overlooked land use category (Padoch et al. 2007). Swiddening or slashing-burning represents the rational use of fire by human beings (Carmenta et al. 2013, 2018), which is a revolution in the history of human evolution and the onset of world civilization (Hough 1932; Stavrianos 1998). The origin of this traditional farming can be traced back to the Neolithic age (Conklin 1961; Mazoyer and Roudart 2007). Furthermore, the swidden system has underpinned and witnessed the slow but continuous development of the Anthropocene (Dumond 1961; Ruddiman 2003). Nowadays, swidden agriculture has been widely practised across the globe with dozens of terminologies, particularly in developing tropical countries (Dressler and Pulhin 2010; Li et al. 2014). Compared to thousands of years of global swidden farming history, the recorded history of swidden agriculture research per se is less than 300 years (Dove 2015a, b), with scattered and sporadic research work in the eighteenth and nineteenth centuries. As the oldest citation database, the Institute for Scientific Information (ISI) Web of Science provides the longest coverage (bibliographic and citation information) from the year 1900 to the present (Li et al. 2010). So far, a thematic review of the history of swidden agriculture research based on this paid-access platform that contains multiple databases back to the beginning of the twentieth century has not yet been reported. The lack of a 20th-century literature review restricts understanding the revival of swidden agriculture research.

Previous review articles on swidden agriculture are characterized by three categories: reviews of regional studies (particularly in the tropics), reviews of one or two specific topics, and comprehensive reviews. Regional reviews are predominantly seen in tropical Asia, including South Asia (Hossain 2011; Kafle 2011; Grogan et al. 2012) and Southeast Asia (Mertz et al. 2009a; Ellen 2012; Li et al. 2014; Dressler et al. 2017). Single-topic and comprehensive essays also examine the studies on vanished swidden agriculture in temperate Asia [mainly Japan (Yokoyama 2013; Yokoyama et al. 2014a; Taisaku 2023)]. On the contrary, review articles on swidden or shifting cultivation in tropical Africa and tropical America are much less reported and old-fashioned (Dumond 1961; Haney 1968; Shukla and Agarwal 1984; Fischer and Vasseur 2000). Topical reviews involve geographic (Schmidt-Vogt et al. 2009), demographic (Mertz et al. 2009b), social (Cramb et al. 2009), economics (Takasaki 2012), and ecological (Mertz 2009) issues at regional to tropical scales over various periods. The frequently reviewed themes include forest recovery and biodiversity (Karthik et al. 2009; Delang and Li 2013), remote sensing and GIS-based mapping (Bhat et al. 2022; Jiang et al. 2022; Mathur and Bhattacharya 2023), effects of swidden agriculture in transition (Filho et al. 2013; Dressler et al. 2017), carbon emission and climate change (Hett and Lu 2014), and debates over the sustainability of swidden agriculture (Fischer and Vasseur 2000; Bhuyan 2019; Nath et al. 2022). Finally, comprehensive reviews generally cover three and more topics of swidden agriculture at tropical and global scales (Dove 1983; Thrupp et al. 1997), in particular the drivers and consequences of the changing swidden agriculture (van Vliet et al. 2012; Martin et al. 2023) and alternatives to this farming (Brady 1996; Pedroso-Junior et al. 2009). Thus, there is a pressing need to conduct more review studies on swidden agriculture at different spatial and temporal scales. More importantly, priority attention is needed to discuss a revival of swidden agriculture research in an era of climate change. How has the history of swidden agriculture research developed since the beginning of the twentieth century, especially in the last decades? What external factors promote swidden agriculture research, and how do they work? Answers to these questions help us understand why swidden agriculture research attracts more attention (Li et al. 2022; Nath et al. 2022).

The structure of this review is as follows. Sect. "Introduction" is about the background, significance, and need to review the history and revival of swidden agriculture research. Sect. "Basic facts about swidden agriculture research" provides two parts based on published papers and the literature database. One depicts two recent overestimates of the present and future global swidden agriculture. The other performs a statistical analysis with Web of Science bibliometric data on swidden agriculture research. Sect. "Developmental stages of swidden agriculture research since the beginning of the last century" defines three developmental stages of swidden agriculture research since the beginning of the twentieth century. Sect. "Three aspects for boosting a revival of swidden agriculture research in this century" elaborates on three aspects boosting the revival of swidden agriculture research this century. Here, a new and plausible hypothesis is proposed: 2008 was a watershed year witnessing a revival of swidden agriculture research. We demonstrated the hypothesis from three aspects: climate change policies, free satellite imagery, and economic crises & recessions. The 2008 global economic crisis has promoted a return to swidden agriculture for poverty-stricken shifting cultivators (Feintrenie and Levang 2009; Sayer et al. 2012). The United Nations Collaborative Programme on Reducing Emissions from Deforestation and Forest Degradation in Developing Countries (or the UN-REDD Programme, or later REDD +) facilitates re-discovering the roles of swidden agriculture in carbon sequestration, biodiversity conservation, and livelihoods. The free and open Landsat data policy and open-access to other satellite images contribute to tracking the history of swidden agriculture evolution (e.g., shortened fallow periods) and transformation (e.g., conversion to industrial plantations). Sect. "Concluding remarks and future perspectives" briefly gives concluding remarks and future perspectives. As swidden agriculture is winning its due academic attention after a century of controversy, we ended the review with an appeal for a multi-sided examination of the changing swidden system in this century (Li et al. 2022).

Basic facts about swidden agriculture research

Overestimates of the present and future swidden agriculture

A satellite-derived dataset of global annual Tree Cover Loss by Dominant Driver (TCLDD) from 2001 to 2019 was collected from the Global Forest Watch (GFW). GFW also provides the best available global and annually-updated tree cover loss data (Hansen et al. 2013). The TCLDD data refers to the annual estimates of tree cover loss associated with five major drivers (Curtis et al. 2018), including commodity production, shifting agriculture, forestry, wildfire and urbanization. Data of the drivers for deforestation has a spatial resolution of 10 km, mainly intended for regional to global usage. Both users' and producers' accuracies range from 81 to 94% for all defined drivers except for urbanization, with shifting agriculture at about 93% and 85%, respectively. Shifting agriculture here refers to temporary or permanent forest loss caused by small- and medium-scale agriculture.

The dataset was exclusively applied to show recently peer-reviewed rough estimates of global swidden agriculture via calculating the national average area of shifting agriculture. Statistics show shifting agriculture is practised in over 160 countries and regions globally, nearly 2/3 in the tropics (Fig. 1), again recognised as one of the five major global forest loss drivers (Curtis et al. 2018). Ten tropical countries have a two-decade average of over 100,000 ha, in descending order, i.e., the Democratic Republic of the Congo (DRC, 759,163 ha), Brazil (670,316 ha), Madagascar (195,406 ha), Mozambique, Côte d'Ivoire, Mexico, Angola, Colombia, Tanzania and Peru (121,203 ha). Notably, the DRC and Brazil cleared over one million hectares annually during 2014–2019, with the maximum (1.45 million ha) in 2016/2017.

Fig. 1

National estimates of land occupied by shifting agriculture based on the 2001–2019 Global Forest Watch dataset

However, the criteria for categorizing a grid as shifting agriculture depends on whether it contains clearings showing signs of agriculture or pasture in most recent imagery and past clearings from visible forest or recovered shrubland in historical imagery from 2000 to 2015 (Curtis et al. 2018). The defining scheme overestimated the area and extent of swidden farming from two aspects. Firstly, current and past clearings don’t discriminate the pixels associated with fires from fire-free ones. Secondly, the clearings showing signs of agriculture, pasture, and forest or shrubland don’t consider topographical factors and others exclusive to swiddening. Those clearings observed in the lowlands (say less than 200 m above sea level (asl)) or high mountains (e.g., over 3000 m asl) can hardly be classified as shifting agriculture (Ziegler et al. 2009a). The over-generalized classification system leads to shifting agriculture detected in some temperate countries with larger territories, such as Russia, Canada, China, the US, Kazakhstan, and Australia. The recent overestimates differ from the long-held view that the swidden system has been primarily practiced in tropical hills and mountains in the past decades (Olofsson and Hickler 2008). Unlike the overestimation of global shifting agriculture (Curtis et al. 2018), there is also a relatively pessimistic evaluation of a considerable shrinkage to the end of this century (Heinimann et al. 2017). The low confidence for detecting and mapping approximately one-hectare swiddens at one-degree cell resolution is also an unbiased reality, not to mention its spatio-temporal dynamics worldwide over the coming decades. Therefore, accurate and consistent mapping of tropical swidden agriculture is essential and vital (Jiang et al. 2022).

Web of Science bibliometric data on swidden agriculture research

The literature search was consistently finished (as of September 2, 2024) via the ISI Web of Science Core Collection. One is to use the commonly reported English keywords of “swidden”, “shifting cultivation”, and “slash and burn”. The other is to use the twelve keyword combinations “swidden, shifting cultivation, or slash and burn” and “biodiversity, carbon, Landsat, or livelihoods”, e.g., the “swidden * carbon” or “swidden * Landsat”. We focused on the peer-reviewed journal publications on swidden agriculture from 1900 to 2023, including Article, Review Article, and Data Paper. Document types were excluded for further statistical analysis, including Abstract, Biography, Book, Clinical Trial, Correction, Early Access, Editorial Material, Letter, Meeting, News, Report, and Other.

The total number searched by the three swidden-like keywords was nearly 2400 (Table 1), with approximately 13%, 25% and 37% in the 1990s, 2000s and 2010s, as well as 18% in the first four years (2020–2023) of the 2020s. A noticeable increase in article number is seen in the mid-1990s and onwards, with annual published articles over 40 per year. The rising focus on tropical swidden agriculture research is highly related to climate change (van Noordwijk et al. 2015). Whilst the idea of REDD was formally recognized at the United Nations Framework Convention on Climate Change (UNFCCC) 13th Conference of the Parties (COP13, Bali/2007), the REDD’s seeds were planted in the 1997 Kyoto Protocol (Holloway and Giandomenico 2009). The publications of swidden agriculture have accounted for nearly 85% since 1997. When the topics of biodiversity, carbon, and livelihoods are considered, the corresponding proportions reach 92% ~ 99% in nearly three decades. Remarkably, implementing of the UN-REDD in 2008 has further stimulated the re-examination of swidden agriculture. Swidden-related publications accounted for over 60% from 2008 to 2023, with over 100 papers yearly since 2016. Meanwhile, the proportions of the articles on biodiversity, carbon, or livelihoods hold 71% to 87%. One can see the vast difference in the amount of literature since 2008. In addition, other literature related to the history, trend, and driver of swidden agriculture was applied.

Table 1 Statistics of article number and percentage of swidden agriculture and some important and relevant topics from ISI Web of Science Core Collection during 1940–2023

Developmental stages of swidden agriculture research since the beginning of the last century

Ever since the earlier observations and descriptions of swidden agriculture as a backward farming practice in the temperate climate regions at the beginning of the twentieth century (Kanthack 1908; Elgee 1914; Gillin 1935; Sanchez 1996), the age-old farming system has slowly and steadily developed under pressure and debates (Dove 1983; Li et al. 2014; van Noordwijk et al. 2015), and even a universal and indiscriminate negation (FAO Staff 1957; Dressler et al. 2017; Pham Thu et al. 2020). It has always been accompanied by pessimistic, unfavourable and detrimental narratives, especially in the last century. As a traditional farming system, the rise and fall of swidden agriculture has always been closely related to global initiatives or policies, such as the 1957 FAO’s call to eliminate swidden (Mertz et al. 2009a), the 1994 Alternative to Slash and Burn (ASB) (Smith and Dressler 2017), and the 2008 UN-REDD Programme (Li et al. 2022). Since the beginning of this century, a shift from almost complete negation and seeking alternatives (Brady 1996; Palm et al. 2005) to re-examination and rediscovery and appealing sustainable development is taking place silently (Li et al. 2014), although the cognitive transition of decision-makers (e.g., government officials) at varied levels toward swidden agriculture still lags behind the scientific community (Dressler et al. 2020; Pham Thu et al. 2020).

Temporally, the fieldwork and research of swidden agriculture mainly undergo three developmental stages (Jiang et al. 2022), i.e., descriptive transcription (Phase I, documenting the history and facts and facilitating cross-comparisons), critical analysis (Phase II, identifying the collective problems and imagining solutions), and comprehensive analysis (Phase III, rediscovering the rationality and sustainability and developing action plans). Grounded on the literature review and differences in annual journal publication numbers, 1957 and 2008 were tentatively identified as the watershed years for the first, second and third phases, even though there may be some subordinate time nodes (Jiang et al. 2022). In 1957, the Food and Agriculture Organization (FAO) of the UN initiated a formal appeal “to governments, research centers, associations and private persons who are in a position to help” “solve this problem in all interested countries” based on “a limited number of case studies” (FAO Staff 1957). The problem here refers to agricultural production's short- or long-term efficiency. By 2008, at least two major influencing aspects accelerated swidden agriculture’s extensive and in-depth research, including cognitive transformation and capacity building. Firstly, the launch and implementation of the UN-REDD Programme as a top-down strategy draw attention back to research the changing swidden system for coping with climate change and adapting to it (Mertz 2009; Li et al. 2014, 2022). Secondly, a new Landsat Data Distribution Policy was released in January 2008. Subsequently, the free access to Landsat and other Landsat-like imagery has fundamentally enhanced the basis of satellite imagery (Woodcock et al. 2008). It has greatly improved our ability to detect and map swidden agriculture and monitor its changes (Jiang et al. 2022).

Phase I (before 1957) is marked by empirical, descriptive, and documentary studies using traditional research methods, including observation, investigation, interview, and/or recording at the field- and household level (Jackson et al. 1950; Schlippe 1956; Conklin 1957). Early studies involved different social science disciplines, such as anthropology, agronomy, ethnoecology, and linguistics. Researchers in this phase usually carried out time-consuming and exhaustive fieldwork to gather thorough knowledge about the full process of traditional swidden practice and fallows. The available small-scale and scattered studies often applied qualitative methods. Among them, American anthropologist Conklin’s (1954, 1957) arduous and pioneer anthropologic field research on Hanunóo agriculture in Mindoro of the Philippines during 1947–1955 was most notable (Dove 2017). He dared to challenge the eradication of shifting cultivation in his era (Ellen 2016; Dove 2017). By contrast, annual counts of academic studies (in English) of swidden, shifting cultivation, or slash and burn were small in this phase. For example, due to technological limitations, very little is known about the geographic and demographic distribution of swiddens and swiddeners at national to global scales. Until the first half of the twentieth century, swidden agriculture's environmental effects and sustainability did not draw much attention as it co-exists with a low world population (Jarosz 1993).

In contrast to sporadic and short-term field studies in the first phase, Phase II (1957–2007) is characterized by gradually systematic scientific research because of the FAO’s appeal, especially since the 1990s. Phase II can be further divided into two sub-phases, with 1997 as a turning year witnessing a transition from complete negation to stirred up interest. Although negative perspectives toward shifting cultivation could be dated back to the 1920s (Zon and Sparhawk 1923), the earlier semi-formal appeal to criminalize and curb shifting cultivation by the FAO only began in 1949 (Pendleton 1950). The negative attitude toward this farming system continued throughout the 1950s (FAO Staff 1957), which heralded the first call to limit or impede global shifting cultivation. Since then, a generally unfavourable view toward swidden agriculture dominated, especially in the first period (1957–1996), although Conklin’s bottom-up investigation on Hanunóo agriculture was remembered as “the single most important critique” (Dove 2017). Albeit with the FAO’s call, or in a context of prevailing opposition, ethnographic, ethnoecologic and ethnobotanical studies of shifting agriculture conducted worldwide in the 1960s (Nye and Greenland 1960; Spencer 1966), 1970s (Johnson 1974; Brush 1975), and 1980s (Peters and Neuenschwander 1988) moved tardily, and social and biophysical scientists seldom worked together (Sanchez 1996).

However, an increase in world population growth till the 1970s (Horiuchi 1993), with three billion in 1960 and four billion in 1974, has further sped up the belittling of swidden agriculture and seeking alternative development. This campaign as a global imperative was more prominent in the 1990s (Brady 1996) and even in the first decade of the 2000s (Zhang et al. 2002) because of the continuous growth of the global population, i.e., five billion in 1987, six billion in 1999, and seven billion in 2011. Specifically, the Alternatives to Slash and Burn (ASB) was first established in 1994 to reduce its threat to humid tropical forests (Pollini 2009) by exploring viable and profitable land use alternatives. Since 2008, it has evolved into the Partnership for the Tropical Forest Margins (Tomich et al. 2007). The change is in response to the UN-REDD Programme aiming at reducing carbon emissions from land cover and land use change (Minang and van Noordwijk 2013), such as agricultural expansion and forest retreat, and supporting livelihoods and ecosystem services (Mertz et al. 2021). Later, an important special issue, “Alternatives to Slash-and-Burn Agriculture”, was launched by Agriculture, Ecosystems & Environment, which pushed the ASB to the forefront of the international research agenda (Brady 1996). The world’s population has surpassed the eight billion mark by 2023, so the ASB’s endeavour is still on its way (Rahman et al. 2017).

The second sub-phase (1997–2007) is a decade of awakening, in which published literature on swidden agriculture increased significantly (Table 1). Despite nearly forty years of eradication efforts, one still does not see the demise of swidden farming in the tropics (Padoch et al. 2007; Ellen 2012; Li et al. 2024). The negative attitude toward swidden farming reaches its peak in the 1990s. Yet, some forward-looking scientists began to shed new insights into the neutral and/or positive effects or benefits in natural and socioeconomic aspects, including biodiversity conservation (de Jong 1997), economic viability and sustainability (Montagnini and Mendelsohn 1997; Vien et al. 2006), and cultural identity maintenance (Schmidt-Vogt 1998). The most representative piece was a World Resources Institute (WRI) Report which discussed the perception change towards eight common myths about shifting cultivation (Thrupp et al. 1997), representing a public concern at that time (Fox et al. 2000; O'Brien 2002; Nielsen et al. 2006).

Phase III (2008 and beyond) has continuously witnessed a long-overdue revival of swidden agriculture research, with journal publications accounting for over 60% in less than 20 years (Table 1). It emphasizes the neutral to positive roles of swidden agriculture in the tropics, such as stabilizing forest cover (Robichaud et al. 2009), transforming landscapes (Ziegler et al. 2011), and the roles of fallow periods in regenerating vegetation and conserving biodiversity. The UN-REDD Programme, delivering on the Paris Agreement and the 2030 Agenda for Sustainable Development, supports over 60 partner countries across the (sub) tropics to protect forests and achieve climate and livelihood goals (Fox et al. 2014; Li et al. 2014). The implementation of the UN-REDD Programme, to a great extent, has thus re-pushed this age-long and marginalized farming system to the foreland of decadal climate change debates (Mertz 2009; Hurtt et al. 2011). Later, the United Kingdom's 2008 Climate Change Act began legalising low-carbon policy (Lockwood 2013). Afterwards, more countries have developed and/or passed climate change legislation (Nachmany et al. 2014). Meanwhile, the UN has successively declared a series of International Years of Biodiversity (2010), Forests (2011), Family Farming (2014), and Sustainable Mountain Development (2022). These and other international initiatives, including the Bonn Challenge and the Sustainable Development Goals (SDGs), further intensify the concerns of swidden agriculture (Li et al. 2022) and function as a top-down strategy for better protection and management of tropical forests (Swamy et al. 2018).

Three aspects for boosting a revival of swidden agriculture research in this century

The year 2008 was a watershed year for swidden agriculture. The research was demonstrated from three aspects, including the global financial crisis, the emergence of UN-REDD, and the free availability of the Landsat archive occurrence in the same year (Fig. 2). Specifically, the worldwide economic crisis have exerted pressure on the livelihoods (e.g., hunger and poverty) of millions of tropical swiddeners. It further challenges the fulfilment of some SDGs in tropical developing countries. The UN-REDD Programme to reduce forest loss and degradation directly challenges the sustainability of swidden agriculture. Finally, Landsat and other finer free-and-open satellite data (e.g., Sentinel-2) is going to facilitate the re-construction of over half-century and pan-tropical geographical and demographic information of swidden agriculture, which is pressingly needed for carbon emission estimation and other related environmental studies.

Fig. 2

The year 2008 as a watershed year for swidden agriculture research

Economic crises and recessions since 2008 increase the uncertainty of the global market and the resilience of swidden agriculture

The year 2008 is increasingly remembered as a watershed year for the global financial and economic system since the Great Depression (Mayer-Foulkes 2010). Crisis gives rise to vulnerability and uncertainty and also gives birth to opportunity. The global financial crisis has brought at least two-sided potential impacts related to the topical forest sector (Ehrenberg-Azcárate and Peña-Claros 2020) and swidden systems, i.e., livelihoods and food insecurity (Vilar-Compte et al. 2015) and export commodity prices plummet and decline (Khoon and Lim 2010). Meanwhile, many countries of the “global south” see a new wave of land grabbing after the 2008 global crisis because of robust demand for food and raw materials, bio-fuels and increased oil prices (Margulis et al. 2013; Costantino 2016).

The global economic crisis signified the severity of global food insecurity due to the 2007–2008 world food price crisis (Vilar-Compte et al. 2015). To ensure food security and livelihoods, shifting cultivators in the tropics have reverted to slash-and-burn farming (Feintrenie et al. 2010; Sayer et al. 2012; Ehrenberg-Azcárate and Peña-Claros 2020), enhanced swiddening intensity (Wunder et al. 2021), or shortened fallow periods (Kilawe et al. 2018). Earlier estimates showed that food-insecure farmers reach nearly 30% in developing countries (Brown and Funk 2008). This figure can be much larger in the tropics, especially during the COVID-19 pandemic. For the ethnic minority groups in the hilly and mountainous areas of the tropics, the swidden system remains an effective way to ensure food (e.g., rice) security (Fu et al. 2003; Setyawan 2016). Due to the 2008 economic crisis, the vulnerability and uncertainty of food acquisition and provision may be more serious (Seguino 2010) as shifting cultivators are more interested in economic returns (Feintrenie et al. 2010). Meanwhile, as tropical forest loss strongly links global commodity (e.g., rubber) markets (Grogan et al. 2019), a plummet of commodity prices limits the export of commercialized forest and agricultural products (Feintrenie and Levang 2009). It thus is likely to cause a slow-down swidden-to-plantation conversion and a return to this buffering livelihood farming system, especially for upland small-scale households. For example, the decline in global demand for natural rubber latex has dropped from the all-time high price of rubber since 2011, which has caused a reduction of rubber plantations (Hurni and Fox 2018).

UN-REDD and other climate policies facilitate re-discovering the roles of swidden agriculture in carbon, biological diversity, and livelihoods

Apart from the impacts of policy and market fluctuation, it also exposes the shifting cultivation to climate variations (Brown and Funk 2008). The swidden system is susceptible due to its dependence on the local climate (Yokoyama et al. 2014b), e.g., sun/wind drying during the peak of the dry season or crop cultivation before rainfall. Therefore, this agriculture is both at risk from and a driver of climate change (Cairns 2015). Due to the close relationship with tropical forests, swidden agriculture is inadvertently pulled back to the central topics of climate change, biodiversity, and livelihoods. At the beginning of this century, scholars start to notice the positive effects of swidden agriculture in following aspects, e.g., climate change mitigation (Mertz 2009; Fox et al. 2014), biological diversity conservation (Padoch and Pinedo-Vasquez 2010), ecosystem services provision (Mertz et al. 2021), and livelihoods improvement (Cramb et al. 2009). For example, according to the dual special issues of Human Ecology focusing on swidden agriculture in the tropics (Mertz et al. 2009a, 2013), they raised several unsettled questions: (1) where swidden agriculture is located, (2) how many shifting cultivators are engaged, (3) how did it evolve and/or it was transformed into another land-use system, (4) how did the length of fallow periods change, and (5) what are the potential socio-economic impacts and biophysical effects (Padoch et al. 2007; Schmidt-Vogt et al. 2009; Mertz et al. 2009b). Later, another special issue discusses and analyzes traditional knowledge of shifting cultivation (Parrotta et al. 2016). However, there still is a surprising lack of spatial and demographic information, let alone their geographical and temporal dynamics. It further adds significant uncertainties for future greenhouse gas emissions estimation (Hurtt et al. 2011), poverty alleviation, and livelihood improvement (Cramb et al. 2009).

The roles of swidden agriculture in carbon budgets

The full cycle of swiddening-cropping-fallowing, ranging from less than ten years to several tens of years, is frequently accompanied by disturbance, damage, and restoration of herbaceous and woody vegetation. The changes in terrestrial vegetation associated with agriculture and forest have a close relation with a constant state of flux of under and above ground carbon budgets (van Noordwijk 2002; Zhang et al. 2002) and affect the fluxes of carbon exchange between the atmosphere and the land (Stocker et al. 2014). Earlier studies have showed that swidden fallow secondary forests contribute to carbon storage (De Jong et al. 2001). However, the carbon sink of freshly opened swiddens and various forms of fallows usually fail to account (Padoch et al. 2007). Although there is missing information on the extent and dynamics of swidden agriculture, this traditional system was not considered in carbon-climate projections models for the Intergovernmental Panel on Climate Change (IPCC) until the early 2010s (Hurtt et al. 2011; Stocker et al. 2014). It should be mentioned that many unknowns of the nexus of climate change and shifting cultivation in transition are to be unravelled (Porter et al. 2019).

Since knowledge of carbon stocks in tropical ecosystems, particularly swidden agricultural systems, is lacking (Mertz 2009), the UN-REDD Programme has greatly drived the carbon-based analysis of swidden agriculture in the tropics (Ziegler et al. 2012; Chan et al. 2016). For example, the WOS-based publications with the keyword combination swidden-like terms and carbon between 2008 and 2023 accounted for 70% of the totality since the 1950s. Some common understandings of the roles of swidden agriculture in carbon budgets are clear. As an example, the evolution of swidden agriculture, including the enhanced intensity of slash-and-burn and a reduction of fallow periods and the transformation into other monoculture plantations, are accompanied by varied-degree decline in below- and above-ground carbon stocks (Bruun et al. 2009). Meanwhile, swidden fallow secondary forests are substantial carbon sinks and their capacity to store carbon is linked to the length of fallow age (Mukul et al. 2016). Therefore, compared with rubber or oil palm plantations, swiddening is considered carbon–neutral or even carbon positive (Bruun et al. 2013, 2018; Fox et al. 2014). It can be seen that the significance of swiddening on carbon sequestration is gradually recognized. However, existing studies usually pay attention to carbon storage within a given stage or the transitional stages, e.g., the contribution of swiddening, fallowing (especially those less than five years) or swidden-to-rubber conversion to carbon emission or fixation (Eaton and Lawrence 2009), while total carbon budgets within a complete cycle are poorly understood (Mukul et al. 2016). In addition, the long-term impacts of shifting cultivation on carbon emission and sequestration are often quantified at the local scale, which further affects global carbon budgets (Lawrence et al. 2010).

The roles of swidden agriculture in biodiversity conservation

Nowadays, anthropogenic footprints in the tropics prevail except for in some reserves. Swidden practice is re-shaping and/or transforming the agricultural and forest landscape (Ziegler et al. 2011). The concern about the negative impacts of the pioneer swiddening system on below- and above-ground biodiversity usually relates to deforestation and forest degradation (de Jong 1997), as it causes the decline in plant and animal species. In the long run, say over hundreds of years, the impacts of swidden systems on forest disturbance can be negligible (Setyawan 2016). In contrast, most established swiddens, which account for a larger proportion with an acceptable fallow length (e.g., over 20 years), can also be biodiversity-friendly (Vien et al. 2006; Padoch and Pinedo-Vasquez 2010). The massive challenge of conserving biodiversity doesn’t originate from the swidden or fallow itself (Russell 1988) but from the transformation into other profitable land use types because of population growth (Rerkasem et al. 2009; Robiglio and Sinclair 2011), such as monoculture (e.g., banana) and industrial timber and pulpwood plantations (Kumar Pandey et al. 2022). There is a voice that swidden agriculture should be recognized for its potential contribution to resource management and agro-biodiversity at the end of Phase II (de Jong 1997; Guo et al. 2002). The termination of swidden agriculture and loss of access to forests of many kinds may cause a collapse of agro-biodiversity (e.g., crop diversity and varieties) (Guo et al. 2002) and further conversely impact the transfer of indigenous knowledge to the next generations and the protection of some essential species (Fu et al. 2003).

Due to the trade-offs between forest resources management and biodiversity conservation (Gupta 2000; Feintrenie et al. 2010), debate on the roles of swidden agriculture in biodiversity tends to continue (Finegan and Nasi 2004). This has a close connection with the understanding of the two parts. On the one hand, tropical swidden production systems comprising various practices are grossly simplified (Finegan and Nasi 2004; Namgyel et al. 2008). It thus easily causes a different understanding of the impacts of swiddening, cropping, and fallowing on biodiversity. On the other hand, existing studies on swidden biodiversity remain in a specific context and do not cover the full range of biodiversity (Norgrove and Beck 2016; Mertz et al. 2021). They usually focus on limited flora and fauna or species diversity in a given area while ignoring the effects on genetic diversity, ecosystem diversity, landscape and cultural diversity (Guo et al. 2002; Padoch and Pinedo-Vasquez 2010). Effects of swiddens and fallows on biodiversity can be diverse from place to place due to the variety of swidden systems and the complexity of biodiversity (Finegan and Nasi 2004). Therefore, there is a need to distinguish site- and species-specific effects on biodiversity related to swidden agriculture systems (Namgyel et al. 2008; Norgrove and Beck 2016).

The critical factor for biodiversity conservation versus shifting cultivation is the length of fallow periods (Gupta 2000). Fallow vegetation and forests of varied ages represent a plant species pool with considerable potential for conserving biodiversity (Robiglio and Sinclair 2011). However, the REDD Programme and government policies on forest and biodiversity further tighten access to open more plots for swiddens, which shortens the length of fallow periods in the face of food insecurity (Setyawan 2016; Kumar Pandey et al. 2022). On the whole, the trend of a reducing fallow and the United Nations Decade on Biodiversity (2011–2020) not only contribute to the implementation of the Strategic Plan for Biodiversity but also promote the systematic studies of biological diversity vis-à-vis swidden agriculture (Li et al. 2014). The publication number searched by swidden-like terms and biodiversity since 2008 accounts for over 80%. Since the evolution of swidden agriculture and the shrinkage of fallow cycles represent the general trend (Kumar Pandey et al. 2022), concern for the conservation of biodiversity could be enhanced in the context of swidden agriculture in transition (Robiglio and Sinclair 2011; Mertz et al. 2021). Also, examining the roles of swidden farming in maintaining and conserving biological diversity should be enlarged and enhanced (Castella et al. 2013).

The roles of swidden agriculture in rural livelihoods

Swidden agriculture has endured for generations due to its role in supporting livelihoods and ensuring food security (Vien et al. 2006; Li et al. 2014). This practice remains essential for the subsistence of indigenous communities from various ethnic groups, such as the Khmu and Hmong, particularly in tropical highland regions above 200 m asl (Mertz et al. 2009b; Pham Thu et al. 2020). These communities typically reside or are resettled in humid tropical regions characterized by acidic, infertile soils, limited access to infrastructure, and low agricultural productivity, such as the Southeast Asian Massif (Scott 2009). They have adapted well to the unique biophysical environments, which turns swidden agriculture into the most suitable production practice or a sustainable one for a slow and stable population for a long time (Russell 1988). This is especially true and necessary when they must cope with climate disasters (e.g., floods and droughts) and/or adapt to socio-economic shifts (Castella et al. 2013). For subsistence purposes, anthropologists and ethnographers have not considered documenting various crops, livestock, and edible or medical plants provided by swidden systems (Conklin 1957; Kumar Pandey et al. 2022).

As economic globalization and regional integration move ahead, national policies on land, economy and access to market and infrastructure have accelerated agricultural expansion and intensification, largely by replacing lands featured by traditional swidden/fallow systems (Jakobsen et al. 2007; Ziegler et al. 2009b). The conversion from either natural forests or swiddens/fallows to industrial plantations (e.g., rubber, eucalyptus, and Acacia mangium) aims to increase the household’s income (Thongmanivong and Fujita 2006) as well as forest cover (Jakobsen et al. 2007). It puts the livelihoods of forest dwellers or shifting cultivators in a more vulnerable stage (Feintrenie et al. 2010; Castella et al. 2013). On the one hand, the visible income comes at the cost of food insecurity, loss of other livelihood options, and cultural identity (Dressler et al. 2017). On the other hand, commercial production of cash crops and profitable timber and pulpwood trees will elevate the demand for agricultural land privatization and grabbing (Thongmanivong and Fujita 2006; Kenney-Lazar 2012). The emergence of worldwide land grabbing exerts unprecedented pressure on the peasants’ livelihoods (Costantino 2016; Kenney-Lazar 2018). The evolution and transformations of swidden agriculture are thus shaking livelihood roles (Jakobsen et al. 2007). Along with the transition in livelihoods, the cultural identity of shifting cultivators is also changing (Dressler 2005). The new generations of swidden farmers are attracted by the external world, full of economic opportunity (Thongmanivong and Fujita 2006), despite the policy uncertainty and the market economy. The sustainability of swidden agriculture is challenged by another critical factor: young generations neither know how to slash and burn nor wish to continue traditional farming (Kumar Pandey et al. 2022).

The international society attempts to achieve the 2030 SDGs, including poverty elimination, zero hunger, and good health and well-being of indigenous people of various ethnic groups. However, the accurate population of global shifting cultivators, especially in the tropics, is still missing (Mertz et al. 2009b), let alone the demographic dynamics. The UN-REDD Programme and other initiatives not only trigger attention to the biophysical effects of a changing swidden system but also reveal the uncertainty of the livelihoods of shifting cultivators. Similarly, the WOS-based results searched by the combination of swidden-like terms and livelihood(s) from 2008 and onwards account for over 87%.

Landsat and other open-accessed imagery contribute to tracking the trajectories of swidden agriculture in transition

Swidden plots are usually scattered in remote hills and mountains, yet remote sensing techniques make them observable without strenuous and repetitive field work (Li and Feng 2016; Li et al. 2018a). Satellite imagery provides a robust and continuous data basis and techniques for detecting and mapping terrestrial changes (Hurni et al. 2013). Among them, the launch of Landsat Multispectral Scanners (MSS) in the 1970s and its successors gave an all-new angle of repeated observing from space and automatic mapping of swiddens and fallows on a computer screen for the first time (Conant and Cary 1977; Bruneau and Le Toan 1978; Cary 1979). So far, the earlier manually sketched or estimated maps of global shifting cultivation appeared in a 1980 book on Economic Geography (Hurtt et al. 2011) and a WRI’s report (Thrupp et al. 1997). Then, 1-km maps of shifting cultivation in South and South East Asia, China, Africa, and South America were generalized using the SPOT-VGT-based Global Land Cover 2000 (Silva et al. 2011). The small size of swiddens per se makes the accuracy of coarse resolution mapping unpredictable or, at best, speculative (Heinimann et al. 2017; Li and Yang 2022). Moreover, a complete picture (e.g., location and extent) of tropical shifting agriculture and its past dynamics since the 1950s’ appeal of eradication, in particular, are still up in the air.

In general, the current progress of remote sensing of swidden agriculture moves slowly, and the development of robust methods is still in progress (Heinimann et al. 2017; Jiang et al. 2022). Technical and methodological limitations include several aspects. First, swidden plots are usually spatially scattered and temporally unfixed. The small-sized nature (Padoch et al. 2007; Jakovac et al. 2016), or about an average size of about one hectare, makes Landsat 30 m imagery (a pixel of 0.09 hectares) one of the most appropriate data sources in the past decades (Li and Feng 2016; Jiang et al. 2022). Second, the temporally dynamic and spatially-diverse features of swiddens and fallows display a complex landscape of agriculture and forest (Padoch et al. 2007). This means that using single-date images to monitor a farming system, spanning from slashing, drying, burning, cropping and fallowing, is impractical (Li et al. 2018a). Third, the tropical settings (e.g., cloud) usually limit the potential of optical satellite imagery. It only makes the mapping of swiddening possible during the dry season (Li et al. 2018b; Yang and Li 2023). Finally, central governments of tropical nations always have contempt for this farming system and don’t take it for granted (Pham Thu et al. 2020). Thus, it is still not considered in cadastral surveys or detailed maps of land use (Schmidt-Vogt et al. 2009). In addition, the commercialization of Landsat in the past makes the annual maximum mapping of newly-opened swiddens and fallows or reconstruction of the trajectory of swidden dynamics unaffordable using historically available scenes, say 600 US dollars per scene (Green 2006).

However, the free access policy of Landsat imagery launched in 2008 encourages detecting and mapping anthropogenic and natural changes in the context of intertwined effects of the growing population and climate change (Woodcock et al. 2008). This is also very timely and indispensable for mapping the evolution and transition of swidden agriculture (Dutrieux et al. 2016; Chen et al. 2023). It allows the academic community to explore robust and expandable algorithms for identifying and mapping shifting cultivation (Li and Feng 2016; Jiang et al. 2022). One apparent phenomenon is the emergence of historical mapping of freshly-opened swiddens, various-age fallows and cropping circles using all available Landsat observations at an annual scale (Dutrieux et al. 2016; Li et al. 2018a). The spatially explicit information of annual freshly opened swiddens and fallows at various ages is essential for estimating carbon stock (Hurtt et al. 2011), biodiversity conservation (Robiglio and Sinclair 2011), and ecosystem services evaluation (Dressler et al. 2017). Making these temporally continuous and spatially explicit studies was challenging before the 2008 free and open Landsat data policy. Statistics show that the WOS-based publications searched by the combination of swidden-like terms and Landsat account for over 68% during 2008–2023.

Meanwhile, the 2010 launch of Google Earth Engine (GEE) at COP-16 in Cancun exerted the advantages of making remote sensing and satellite imagery accessible (Kelley et al. 2017). It transits expensive satellite imagery and extensive computer processing to the cloud, facilitating complex satellite analysis worldwide, even in developing countries, and bringing remote sensing and satellite data one step closer to the policy world. Assimilation of satellite data, especially Landsat, the longest continuous space-based imagery, into biophysical process models would be promising for studying carbon and biodiversity (Inoue et al. 2010; Kelley et al. 2017). Similarly, the European Copernicus Program and other Earth Observation satellites adopt a free and open policy. It can also be expected that the harmonization of Landsat and Sentinel-2 employing GEE will generally help to improve the detection of small, subtle, frequent, dynamic, diverse and complex changes in swidden agriculture in the humid tropics (Claverie et al. 2018; Li et al. 2024). In addition, by combining the datasets of global gridded population (e.g., LandScan), the issues of demography and the changes involved with shifting cultivation can also be estimated shortly (Mertz et al. 2009b).

Concluding remarks and future perspectives

Swidden agriculture is evolving by shortening the length of fallow periods and/or being transformed into other commercial tree plantations (Ziegler et al. 2009b), such as rubber, teak, and pulpwood trees, in the context of growing population, climate change, and the continuous economic recession since 2008. However, it will persist in the foreseeable future (van Vliet et al. 2012; Li et al. 2014, 2022; Nath et al. 2022). Meanwhile, we believe that this traditional farming system will gain more attention and ongoing controversy, especially in climate change mitigation (Mertz 2009; Fox et al. 2014) and biological diversity conservation (Padoch and Pinedo-Vasquez 2010).

The major takeaway from peer-reviewed journal articles is that 2008 marked a pivotal year for reviving research on swidden agriculture in the tropics following a century of debate. Several key factors contributed to this resurgence. First, the global economic crisis profoundly impacted the livelihoods of forest dwellers, especially shifting cultivators, leading to challenges like food insecurity and the erosion of cultural identity. The crisis limited the ability of small-scale households to export commercialized agricultural and forest products while also triggering a new wave of land grabbing by local cooperatives and foreign enterprises in the global south. Second, initiatives like the UN-REDD Programme and other environmental efforts prompted scientists and policymakers to revisit the known and potential benefits of traditional swidden systems. These systems are now recognized for enhancing carbon storage, conserving biodiversity, and providing vital ecosystem services. Additionally, they contribute to stabilizing and improving livelihoods while preserving ethnic beliefs. Furthermore, the 2008 UK Climate Change Act, the first legally binding national framework to address climate change, brought carbon-related swiddens and fallows back into the spotlight. International and national climate policies reignited debates about the role of swidden agriculture in mitigating climate change. Lastly, the 2008 policy change that made Landsat data freely available for the first time empowered scientists with cloud computing capabilities. This access enabled researchers to re-examine important questions about swidden agriculture, such as its current location and extent, past and future transformations, and its evolving impact on the environment and communities. In contrast to widespread anti-swidden narratives, a more balanced and objective approach to swidden agriculture is expected to emerge.