Avaliação de modelos e métodos geoestatísticos para a estimativa de biomassa em área de caatinga

Abstract

After a crisis of 1974 the Caatinga has suffered severe interventions, mainly as a forest suppression for the energetic supply of industries and the local population, neglecting an availability of the phytomass stocks in the diverse phytophysiognomies of Caatinga, mainly in less favored regions in precipitation, where The size of trees is less developed. Faced with this problem, a research with the objective of adjusting mathematical models to aerial biomass and other compartments of woody plants in the Caatinga, besides verifying an application of geostatistical techniques for a spatial estimation of the biomass in these regions. The research was carried out in Caatinga arbustivo-arbórea, in the municipality of Floresta - PE. For the forest inventory, 35 units of 20mx20m were distributed 80m apart and 50m edges, where all plants with diameter at breast height (DAP≥1.9cm) were measured and submitted to floristic and phytosociological analysis only to identify 5 Species of size VI, in sequence as they were distributed in 5 diametric classes, being chosen 8 classes per class, totalizing 200 shrub-arboreal specimens for the collection of biomass by the destructive method, possession of the data to quantify possible as proportions of biomass by Classes And compartments, as well as mathematical models by the plant and for the total. It was also verified a spatial dependence of tree adaptation and an adjustment of geostatistical models with a base of management techniques and ordinary coordination, using semivariographic and cross validation analysis. (43.2%)> thick branch (26.6%)> fine branch (25.9%)> leaves (4.3%), among the compartments as leaves presented the worst fit with R2 = 0.55, already The Fuste, Total, Thick Branch and Thin Branch express values of R2 = 0.89; 0.86; 0.71 and 0.72 in order, with percentage errors for the equations S (xy)% = 22.3; 26.0; 0.54 and 0.35, the models that best fit those of Schumacher-Hall and Gompertz. For the geostatistical modeling of a technique of detection of spatial dependence, with a range of 17.6 meters for dry biomass, for a spherical model with values of R2 = 0.61 for cross validation, it has already been calculated for R2 = 0.83 For the exponential model with strong spatial dependence for dry biomass (kg), with a range of 10 meters. It is concluded that as proportions of biomass per compartment follows a harmonious distribution pattern, it is possible to generate reliable equations for the parts of a leaf plant and thin branches for a species Commiphora leptophloeos. As geostatistical techniques employed, they captured a strong spatial dependence of biomass and DAP (cm), with good precision in the estimates made in Caatinga areas.