The woodlands in the south-west of the Argentinean pampas are dominated by Prosopis Caldenia Burkart (calden). The current deforestation rate of this woodlands is 0.82% per year. Different compensation initiatives have begun that recognize the role of forests as environmental service providers. The financial incentives they offer make it necessary to quantify the amount of carbon stored in the forest biomass. A model for estimating calden biomass was developed. Thirty-eight trees were selected, felled and divided into sections. An equation system was fitted using joint generalized regression to ensure the additivity property. A weighted regression was used to avoid heteroscedasticity. In these woodlands fire is the main disturbance and it can modify tree allometry, due this all models included the area of the base of the stem and tree height as independent variables since it indirectly collects this variability. Total biomass and the stem fraction had the highest R2_Adj. values (0.75), while branches with a diameter less than 7 cm had the lowest (0.58). Tree biomass was also analyzed by partitioning into the basic fractions of stem, crown, roots, and the root/shoot ratio. Biomass allocation was greatest in the crown fraction and the mean root/ shoot ratio was 0.58. The carbon stock of the caldenales considering only calden tree biomass is 20.2 Mg ha(-1). While the overall carbon balance of the region is negative (deforestation and biomass burning, the remnant forested area has increased their calden density and in an indirect way his carbon sequestration capacity could also be increased. (C) 2014 Elsevier Ltd. All rights reserved.