Assessing land-use and carbon stock in slash-and-burn ecosystems in tropical mountain of Laos based on time-series satellite images

Abstract

In the tropical mountains of Southeast Asia, slash-and-burn (S/B) agriculture is a widely practiced and important food production system. The ecosystem carbon stock in this land-use is linked not only to the carbon exchange with the atmosphere but also with food and resource security. The objective of this study was to provide quantitative information on the land-use and ecosystem carbon stock in the region as well as to infer the impacts of alternative land-use and ecosystem management scenarios on the carbon sequestration potential at a regional scale. The study area was selected in a typical slash-and-burn region in the northern part of Laos. The chrono-sequential changes of land-use such as the relative areas of community age and cropping (C) + fallow (F) patterns were derived from the analysis of time-series satellite images. The chrono-sequential analysis showed that a consistent increase of S/B area during the past three decades and a rapid increase after 1990. Approximately 37% of the whole area was with the community age of 1–5 years, whereas 10% for 6–10 years in 2004. The ecosystem carbon stock at a regional scale was estimated by synthesizing the land-use patterns and semi-empirical carbon stock model derived from in situ measurements where the community age was used as a clue to the linkage. The ecosystem carbon stock in the region was strongly affected by the land-use patterns; the temporal average of carbon stock in 1C + 10F cycles, for example, was greater by 33 MgC ha⁻¹ compared to that in 1C + 2F land-use pattern. The amount of carbon lost from the regional ecosystems during 1990–2004 periods was estimated to be 42 MgC ha⁻¹. The study approach proved to be useful especially in such regions with low data-availability and accessibility. This study revealed the dynamic change of land-use and ecosystem carbon stock in the tropical mountain of Laos as affected by land-use. Results suggest the significant potential of carbon sequestration through changing land-use and ecosystem management scenarios. These quantitative estimates would be useful to better understand and manage the land-use and ecosystem carbon stock towards higher sustainability and food security in similar ecosystems.

Introduction

Quantitative assessment of carbon exchange between various ecosystems and the atmosphere is one of the critical subjects in environmental policy-making (IPCC, 2003, Houghton, 2005, Hese et al., 2005, Ramankutty et al., 2007, Fung et al., 2007) as well as in related sciences since the carbon stock in terrestrial ecosystems plays an important role in the global carbon cycle (Tian et al., 2003, Dong et al., 2003, Bellamy et al., 2005). Land-use changes such as conversion between cropland and forest may have great impacts on carbon exchange at the interface of land surfaces to the atmosphere (Houghton and Hackler, 1999, IPCC, 2003, Righelato and Spracklen, 2007, Meyfroidt and Lambin, 2008). Amazonian forest has been focused by many studies (e.g. Cochrane et al., 1999, Houghton et al., 2000, Hirsch et al., 2004, Davidson et al., 2008, Vargas et al., 2008). Nevertheless, estimating the carbon balance in many tropical regions remains an important challenge due to limited data and their uncertainty (Ramankutty et al., 2007).

In the tropical mountains of Southeast Asia, slash-and-burn (S/B) agriculture, i.e. shifting cultivation is an important food production system, and widely practiced in mountainous regions of Vietnam, Laos, China, Bangladesh, Myanmar and northern India (Rasul and Gopal, 2003). In S/B agriculture, a patch of vegetation is cleared, burnt, and used to grow crops for a few seasons, and then abandoned for regeneration of vegetation. Hence, dynamic and large-scale changes in the ecosystem carbon stock occur over the cropping (C) + fallow (F) cycles. It has been pointed out that the S/B agriculture is a rational crop production system to utilize hilly lands with a little input, and used to be sustainable while the fallow period was long enough to recover the plant biomass and soil fertility (Fujisaka, 1991, Roder, 2001). The fallow biome also provides various non-timber forest products (NTFP; mushrooms, medicinal plants, bamboo, insects, etc.) that play an important role in rural livelihoods (Douangsavanh et al., 2006).

Nevertheless, increasing population pressure and land-use regulation have forced shifting cultivators to expand S/B areas and to shorten the fallow period (Pravongvienkham, 2004, Douangsavanh et al., 2006, Fujita and Phanvilay, 2008). Consequently, the negative effects on crop productivity such as decreasing soil fertility and increasing weed problems have been reported (Fujisaka, 1991, Roder, 2001), but also the negative impacts on biological resources (forest, NTFS, biodiversity, etc.) and the atmosphere are serious concerns. Therefore, alternative land-use and ecosystem management scenarios are urgently required to improve food and resource security (Douangsavanh et al., 2006) as well as for carbon sequestration. However, quantitative information on the chrono-sequential change of land-use and ecosystem carbon stock in the region is not available. Limited accessibility to the mountainous areas as well as the limited availability of statistical data for the region has hampered scientific investigations, especially at regional scales (Fujisaka, 1991, Roder, 2001, Shrestha and Zinck, 2001). In general, it is difficult to obtain quantitative information especially on land-use history (Aumtong et al., 2009).

Use of satellite imagery is one of the viable options for such investigations. A great deal of efforts have been made to estimate biophysical variables such as biomass, net primary productivity (NPP) and light use efficiency as well as land-use change in a wide range of ecosystems (e.g. Dymond et al., 2001, Dong et al., 2003, Hese et al., 2005, Veroustraete et al., 2002). Spectral assessment of photosynthetic efficiency and capacity is one of vital interests in remote sensing and ecosystem science communities (e.g. Inoue et al., 2008a). Assimilation of remotely-sensed data into biophysical process models would be another promising approach (e.g. Nouvellon et al., 2001, Veroustraete et al., 2002, Inoue and Olioso, 2006, Dorigo et al., 2007). In spite of its usefulness, availability of optical satellite imagery is strongly limited due to the tropical climate conditions in the region and the use of radar imagery (SAR) remains challenging due to the steep topographic conditions. Therefore, the simple approach using time-series satellite imagery examined by Inoue et al., 2007, Inoue et al., 2008b may be feasible for the geo-spatial assessment of land-use and carbon stock in S/B ecosystems under such unfavourable conditions.

Thus, the main objective of this study was to provide quantitative information on the land-use and ecosystem carbon stock in the region by linking time-series satellite imagery with an ecosystem carbon model derived from in situ experiments. Another objective was to infer the impacts of various land-use and ecosystem management scenarios on the potential of carbon sequestration at a regional scale.