Why is shifting cultivation practiced

Fire is one tool that is carefully used in these farming systems, which have been developed over generations to be appropriate to both the land and the community. In the mid to late 20th century, swidden was seen as a disastrous activity which was destroying forests that should either be used for conservation or logging.

Several governments banned or severely restricted the practice whilst allowing vast areas of forest land to be cleared for logging, plantations and biofuel production.

This attitude prevails. In the Peruvian Amazon, scientists have recorded an average of 37 species of tree on swidden plots and records show over species are supported in the territories of the Karen swidden farmers in northern Thailand.

Far from being responsible for destroying biodiversity, tribal swidden systems are being recognised as having contributed to the diversity of forest areas and to maintaining the ecological value of these areas. This shows that tribal peoples are better at looking after their environment than anyone else. The Wanniyala Aetto of Sri Lanka are a forest people who have traditionally practiced a form of agriculture called chena cultivation.

The forest plots are cleared for one or two years and then left to rest for seven or eight. The plots look untidy, with a multitude of different plants coexisting, rather than neat beds of specific crops, but this diversity is the key to both their ecological value and their economic importance to the Wanniyala Aetto.

During the fallow periods, the regenerating plots continue to provide for the community — from animals to hunt to a range of forest products including medicines, fibres and fuel.

In these communities, forests are a valued part of their livelihood system — rather than something to be cleared to make way for fields. The Jummas — a group of several tribes in the Chittagong Hill Tracts of northern Bangladesh — are named after Jhum, their form of shifting cultivation. The Jumma tribes have developed this farming system to suit the rugged, hilly landscape in which they live. These have been criticized because the benefits for biodiversity are unclear. However, this is only one side of the story, as agricultural scientists show in a new study.

The rapid expansion of oil palm The study offers an An international research team questions this. According to the authors, a And researchers Spiders' Web Secrets Unraveled. Even in its short-cycle form, it allows farmers a diverse harvest of 20—30 crops sometimes more than 40 , cultivated simultaneously throughout the year. This mixed cropping allows for year-round food availability and dietary diversity.

It also enables farmers to avert risks and is central to weather mitigation and pest management. The second is its unique land use practices: the sequential alternation of agriculture and fallow forestry on the same plot. This integration effectively neutralizes land degradation, contributes to long-term sustainability, and ensures the resilience of the food system and the farmers. The third is its built-in tenure system. Shifting cultivation functions under the principles of common property, in which ownership remains a community right held by the clan or village and land is allocated to families according to household size.

This system ensures universal access to land and resources in an egalitarian and equitable manner. How can we overcome the challenges facing the revitalization of shifting cultivation? We need to change the narrative surrounding shifting cultivation to one that promotes its strengths. This will require intense, dedicated engagement between development agencies, governments, academia, and indigenous peoples to showcase good practices and formulate strategies for the revitalization of shifting cultivation.

In addition, the methods used in the different national and regional assessments vary greatly and are far from being standardized. As global level ground data collection is not feasible and no global data on shifting cultivation for the considered time span of to is available, we generated a validation dataset, which contains a detailed delineation of the areas under shifting cultivation for a stratified sample of one-degree cells. The stratified validation sample design was chosen considering the distribution of the validation samples per occurrence class and the spatial distribution per continents.

Taking into consideration the stratification criteria, one-degree cells were randomly selected to be validated. As for the entire global level assessment of all the 2, one degree cells considered, the GEFC and available very high resolution imagery i. Bing and Google Earth were used.

In addition, to the extent multi-temporal very high resolution imagery was available on Google Earth, these were used to get an even better impression of the landscape and its land-use dynamics. According to Olofsson et al.

Each of these samples was visually examined in detail at scale of , or lower for the presence or absence of the above mentioned shifting cultivation specific spatio-temporal signs of clearing and regrowth on the landscape. Fig 2 illustrates this process for one sample cell.

In this specific case or Fig 2A : the global distribution of the stratified sample of the one-degree cells used in the validation data set. The red box marks the extent of Fig 2D and Fig 2E.

Geological Survey. We used a search string similar to the one used for the literature review:. The search was performed in September and generated results. After eliminating papers that were not related to recent shifting cultivation in the tropics e. The first authors of these papers were listed and their email addresses were found.

An online questionnaire was designed and sent to these authors in September using survey monkey. All answers were anonymous and cannot be traced back to the individual expert. Thirty-eight email addresses were no longer functional and 7 authors declined to participate for various reasons. Of the remaining authors, 72 responded and 49 provided usable information see S1 File. There was a bias towards responses from researchers who worked in Southeast Asia see Fig 3 ; but this is also the world region where most research on shifting cultivation has been done, whereas Africa has the fewest studies and is clearly under-represented in light of the considerable occurrence of shifting cultivation there [ 17 , 44 ].

This figure was elaborated by the first author using ArcGIS The respondents were asked to estimate the current spatial extent of shifting cultivation as well as past and future trends in the development of this extent in their area of interest.

The most important questions included:. The information provided by the respondents related to very different spatial scales, ranging from village to district, provincial, and, in some cases, national scales. Moreover, it is not possible to know in detail how the respondents arrived at their assessment of past and future extents of shifting cultivation. The selection process ensured that all are experts in the field, but there may be disciplinary or personal differences in the way that especially the future of shifting cultivation was assessed.

With these caveats in mind, we aggregated responses to the national scale and to three supranational regions: the tropical parts of 1 Central and South America, 2 Africa, and 3 Asia. We then made an estimate of trends in occurrence of shifting cultivation for these aggregated regions for , and also taking into account the historical trends between the Butler map and our classification. It is important to note that these are indeed very rough possible scenarios and should be seen as expected trends rather than fixed percentages of decline.

To generate a spatially explicit prediction of the temporal dynamics decline of shifting cultivation through to the s, we combined the survey results with several simple assumptions.

We initiated occurrence at a one-degree resolution in base year at the mean of the above-stated ranges for each occurrence category in the assessment of current landscape with signs of shifting cultivation based on the GFC data c. For each grid cell, the occurrence of shifting cultivation declined linearly by the mid-point of the estimated losses in , and see Section 3.

If the survey provided information about when i. The map is binary presence-absence , with no information on occurrence frequency or land-use intensity. At the regional scale, an assessment by Spencer [ 45 ] provides a slightly more differentiated picture of the presence of shifting cultivation practices in Asia. The data sources are not entirely clear, but the assessment seems to be based on a number of studies carried out between the end of the 19th century and the late s, along with expert judgement.

Spencer [ 45 ] indicates an approximate area under shifting cultivation currently cultivated fields plus all stages of fallows of around million hectares Mha for Asia. The classic studies on shifting cultivation in Africa [ 46 — 48 ] do not provide any area data for the continent, and for Central and South America we were unable to find any regional-scale past area estimation. In the absence of further data to validate past estimations of the extent of shifting cultivation, the Butler map may be considered a reasonable representation of the global distribution of shifting cultivation from into the s.

Although the areas shown on the map also include non-shifting forms of subsistence agriculture, it is reasonable to assume that shifting cultivation predominated in the humid and sub-humid tropics, which are the main focus of the present study. Our review of the more recent literature revealed surprisingly few studies containing regional or global estimates of areas under shifting cultivation. An area of 1, Mha was mentioned by Davidson et al.

The study by Silva et al. This approach, however, suffers from the shortcoming that land cover data are of very limited use in estimating land use practices, which is acknowledged by Silva et al. For example, the large areas of tree plantations established in Vietnam and southern China in the s and s can also leave a mosaic signature at a certain point in time and might therefore be incorrectly interpreted as shifting cultivation Fig 1.

On the other hand, large areas with shifting cultivation in Central Africa, for example in the Democratic Republic of the Congo [ 36 ], are strongly under-represented in Silva et al. The number of estimates of areas under shifting cultivation at regional and national scales in the literature is also very limited. For Southeast Asia, Schmidt-Vogt et al. However, the areas indicated for each country vary greatly depending on the source.

For example, data for Laos indicate between 2 and over 6 Mha, while more recent figures based on remote sensing multi-temporal Landsat for northern Laos—where the largest share of shifting cultivation in the country is found—are 3.

Comparison of these figures with the GLCbased area estimate for Laos of almost 11 Mha [ 1 ] underlines the problem of using the GLC to estimate areas under shifting cultivation.

For India, Goswami et al. Silva et al. As for Africa, we found only one recent national study, on the Democratic Republic of the Congo [ 36 ].

The authors estimated that these areas made up Fig 5 presents the results of our own visual approximation of the global extent of shifting cultivation around at a one-degree resolution, based on Hansen et al. The map shows that shifting cultivation is still present across large areas of the humid tropics. However, the occurrence of shifting cultivation within most of the individual one-degree cells is very low, meaning that it is a minor component of the overall landscape.

More widespread signs of shifting cultivation were found mostly in small pockets, with the exception of larger areas in Central Africa e. Based on visual inspection of annual global deforestation data [ 8 ] from to and very high-resolution satellite imagery. Areas in which shifting cultivation can be assumed to have never existed or disappeared decades ago have been excluded from the analysis dark gray. The remaining cells 3, were excluded from the analysis as shifting cultivation can be assumed to have never existed or disappeared decades ago see Fig 5 and Method section.

The very low occurrence of shifting cultivation within a majority of cells, particularly in the Americas, points towards shifting cultivation being either a form of cultivation practiced in landscapes where only a minor share of the land is used for agriculture e.

The validation of the estimation of landscapes showing signs of shifting cultivation revealed that Also when considering the different levels of occurrence of shifting cultivation estimated based on the overall impression of the landscape per one-degree cell, the accuracies were high see confusion matrix in Table 2 , with an overall classification accuracy of This indicates that, despite the subjectivity involved on estimating the landscape level of occurrence of shifting cultivation in our classification, the method led to reproducible and accurate results.

While an area approximation of actual shifting cultivation landscapes based on our analysis is difficult due to the estimated and not measured shared of shifting cultivation of each one-degree cell we believe that, given the high accuracy of this estimation see Table 2 , a conservative estimate can nonetheless be derived by visually inspecting the cells and allocating reasonable shares of shifting cultivation landscape currently cultivated fields plus all stages of fallows to them as described in the methods section.

Using the mean values of the ranges specified for the different occurrence classes for all cells in all classes results in a total area of Mha. This number certainly requires further validation before it can be claimed to be an adequate estimate of the global area under shifting cultivation.

Nonetheless, it is more evidence-based than the 1, Mha of unclear origin that are repeatedly cited in the literature e. Surprisingly, our estimate is not too far from the Mha proposed in Silva et al [ 1 ], even though their estimate excludes large areas under shifting cultivation and includes areas under other forms of agriculture and natural vegetation.

At the regional scale, the only number to which we can compare our result is the Mha for Asia estimated by Spencer [ 45 ], which is considerably larger than our estimate for this region of approximately 70 Mha.

Two steps enabled us to gain insights into larger recent last 40 to 50 years trends in the development of the global area under shifting cultivation.

First, we compared the Butler [ 13 ] map Fig 4 with our own spatial estimate Fig 5 of areas that were under shifting cultivation around Second, we combined the results of this comparison with those of our expert survey.

The difference between the Butler map—which is presumably based on studies dating from the s and s [ 13 ]—and our own current investigation around is displayed in Fig 6. Generally, the two approximations of areas under shifting cultivation are in fairly good agreement, especially when considering that the large differences in the arid areas of northern Africa, southern Angola, and Zambia and the high-mountain areas of eastern Bolivia are largely explained by the fact that Butler included other, non-shifting forms of extensive smallholder agriculture and that shifting cultivation in these areas based on the climatic condition never could have been widespread.

This is the case for Peninsular Malaysia [ 28 ], parts of Sumatra [ 60 ], Yunnan Province in southern China [ 61 ]and the southern part of the Brazilian Amazon [ 62 ], for example. For this reason, the known widespread decrease in the occurrence of shifting cultivation in Southeast Asia e. Laos, Vietnam over the past 15 to 25 years is not reflected in this comparison. The areas newly classified as having shifting cultivation blue in Fig 6 are more likely to have been missed in the Butler map than to represent actual new areas under shifting cultivation, as many of these regions are well known to still have significant shifting cultivation landscapes.

Such areas are found in parts of Southeast Asia e. Myanmar [ 63 — 65 ] , Central America e. Yucatan Peninsula [ 66 ], northern South America e. Venezuela [ 67 ] , and Madagascar [ 68 ].