Long-term explorations serve to widen the perspectives of decision makers. Biophysical and technical possibilities and constraints are confronted with the valuedriven objectives of stakeholders in Multiple Goal Linear Programming (MGLP) models. Two methodological aspects of long-term explorations are elaborated in this thesis: uncertainty in agro-ecological coefficients and temporal aspects of land use. The effects of these aspects on generated land use scenarios are studied using data from the Northern Atlantic Zone of Costa Rica (NAZ).Uncertainties in agro-ecological coefficients concerning nutrients and biocides were quantified. Only uncertainties caused by lack of knowledge of underlying biophysical processes or lack of data for quantification were considered. "Average", "pessimistic" and "optimistic" estimations of coefficients were generated, based on different perceptions of the influence of environmental factors. The estimations of the coefficients for various production activities are strongly correlated owing to the assumption of "best technical means" (i.e. inputs are used with the highest technical efficiency according to available knowledge and techniques). These coefficients were used in the single-period MGLP-model that was constructed for the NAZ. With the help of sensitivity analyses the effect of uncertainties on land use scenarios was determined for five tentative policy views, representing different perceptions of sustainability. It is concluded that, in long-term explorations, uncertainties in agroecological coefficients strongly affect the objective function values. However, they hardly affect the optimal land use allocation, because the ranking of production activities for the agro-ecological coefficients hardly changes when including uncertainties.In long-term explorations the following temporal aspects are relevant: 1. Growth and ageing of crops and livestock, 2. Fluctuations in coefficients caused by variation in weather conditions, 3. Interactions in time. After an inventory of ...