When fuelwood is harvested at a rate exceeding natural growth and inefficient conversion technologies are used, negative environmental and socio-economic impacts, such as fuelwood shortages, natural forests degradation and net GHG emissions arise. In this study, we argue that analyzing fuelwood supply/demand spatial patterns require multi-scale approaches to effectively bridge the gap between national results with local situations. The proposed methodology is expected to help 1) focusing resources and actions on local critical situations, starting from national wide analyses and 2) estimating, within statistically robust confidence bounds, the proportion of non-renewable harvested fuelwood. Starting from a previous work, we selected a county-based fuelwood hot spot in the Central Highlands of Mexico, identified from a national wide assessment, and developed a grid-based model in order to identify single localities that face concomitant conditions of high fuelwood consumption and insufficient fuelwood resources. By means of a multicriteria analysis (MCA), twenty localities, out of a total of 90, were identified as critical in terms of six indicators related to fuelwood use and availability of fuelwood resources. Fuelwood supply/demand balances varied among localities from -16.2 +/- 2.5 Gg y(-1) to 4.4 +/- 2.6 Gg y(-1), while fractions of non-renewable fuelwood varied from 0 to 96%. These results support the idea that balances and non-renewable fuelwood fractions (mandatory inputs for Clean Development Mechanism (CDM) cookstoves projects) must be calculated on a locality by locality basis if gross under or over-estimations want to be avoided in the final carbon accounting. (C) 2009 Elsevier Ltd. All rights reserved.