||Keywords: Rainfed lowland rice-based systems, Northeast Thailand, nutrient balance analyses, sustainability assessment, sustainable natural resource management, integrated rural development strategies, livelihood analyses, contextual analyses, multi-scale approach, interdisciplinary approaches, holistic approaches, decision support, Least Developed Countries, Ethiopia, Mozambique.InNortheast Thailand, the sustainability of rainfed lowland rice-based systems, the dominant land-use system (LUS) in the region, is a concern for livelihood development in this relatively poor area of the country. Poor soil fertility and low inputs are considered major causes of the sustainability problems. Similar problems exist for a wide range of LUS in the developing world. Reversal of such developments requires integrated rural development strategies aimed at breaking the environment-poverty downward spiral.The key hypothesis for this thesis is that, efficient and effective, rural-development strategies, aimed at the twin-objectives of sustainable natural resource management (SNRM) and improved and sustainable livelihood development, require 'innovative', i.e. more participatory, integrated and holistic approaches, while following a strict priority setting by combining holistic concepts and a small selected set of participatory, integrated and/or complementary analyses to keep efforts manageable and result-oriented. The challenge therefore is to find the appropriate balance between on one hand the demand for more comprehensive contextual and integrated analyses and on the other hand the requirement to keep such efforts manageable in terms of human , material and financial resources. The key hypothesis was partly tested in a collaborative research and development (R&D) programme inNortheast Thailand. Additional evidence was based on literature study and later experiences, including work in sub-SaharanAfrica. The introductory chapter elaborates relevant concepts, principles and approaches, such as the multiple dimensions of SNRM, the sustainable livelihood concept and approach (SLC/A) and nutrient balance analyses, which are useful for the work in Northeast Thailand, but also - often even more essential - for development strategies in the least developed countries (LDCs) of the world.The collaborative R&D programme in Northeast Thailand explicitly examined the relevance of integrated nutrient budget, soil fertility management and livelihood analyses within their broader - dynamic resource management domain (DRMD) - context. This entailed DRMD characterization, multi-scale partial nutrient balance analyses (PNBA)and integrated environmental and socio-economic analyses, including livelihood and correlation analyses, as well as integrated environmental and socio-economic accounting. Large variations in partial N, P, and K balances were found among and within farms, especially at land utilization type (LUT) level. Although the mean values were positive, many negative partial balances were observed, especially at the LUT level. These large variations are the result of the fact that farmers manage nutrients for similar parcels of land in very different ways, even for LUTs with identical cropping systems within the same farm. Results were assessed with specific emphasis on biophysical aspects, in line with a discussion on methodological aspects. Diversification of income sources, through off-farm employment, non-agricultural on-farm income, such as weaving, and diversification of the agricultural system beyond rice, had a large impact on household wealth. This, in turn, affected the capacity of the household to manage the natural resources of the farm. No significant correlation was found between total income or non-rice income and nutrient inputs; however, this does not mean that they are unrelated. Based on fertilizer use and prices, mean elemental N, P and K retail prices were calculated and used for integrated environmental and economic accounting. Based on mean partial N, P and K balances, partial N, P and K balances were calculated in monetary terms. The results followed the average positive partial balances for rice-based systems established for a sample of 30 farms with large variability among farms and, even more so, among land utilization types, distinguished cropping system-management combinations, however, for each nutrient to a degree depending on its price. Partial nutrientbalances were most extreme for N and K. In contrast, in monetary terms P balances were most extreme. These analyses revealed that, due to the common use of N-P or N-P-K compound fertilizers, investments in P that is generally non-limiting and rather expensive, are too high.Some alternative - prototype - cost benefit analyses (CBA) have been performed for rice cultivation in 1999 on one farm included in the main survey. Complementary and integrated biophysical and socio-economic analyses are relevant for decision support for SNRM and livelihood development, including the development of decision support tools for farmers and extension workers. The collaborative R&D programme served as an exemplary and paradigmatic case for innovative approaches and contributes to the advocacy for their broad adoption.Other relevant aspects are briefly alluded to, including less tangible ones such as governance, decentralization, organizational and institutional development, and capacity development. Broader perspectives are also highlighted based on a summary overview of major challenges and opportunities for rural development inEthiopiaandMozambique, respectively. This is followed by a general discussion that advocates that SNRM can only be achieved by following a livelihood approach, including a strong emphasis on socio-economic incentives. In addition, major challenges are identified for multi-scale sustainable institutional development, which requires efforts aimed at good governance.