ABSTRACT This study aimed to evaluate the spatial dependence between agrometeorological variables and upland cotton yield the microregion of Cariri Oriental, Paraíba state (Brazil), using weighted spatial modeling system. In this study, we used the historical agricultural production data from the Brazilian Institute of Geography and Statistics (IBGE), observed rainfall data from the Executive Water Management Agency of Paraíba State (AESA), and air temperature averages estimated by the Estima_T software. Spatial regression models (classical - CR, autoregressive - SARM, and spatial error - SEM) were used to correlate the dependent variable (upland cotton yield) with covariates (agrometeorological variables). Fitted model parameters were estimated using the Maximum Likelihood method. Model performance was evaluated based on coefficient of determination (R2), maximum likelihood function logarithm (AIC), and spatial residue analysis. Moreover, an exploratory spatial analysis allowed us to verify spatial autocorrelation between upland cotton yield and agrometeorological elements, using statistical tools such as Moran’s index I.

RESUMO Diferentes tipos de vegetação estão distribuídos em montanhas de acordo com a altitude, topografia e solo. A composição e estrutura das comunidades de aves nestas áreas podem mudar com o gradiente de vegetação, com comunidades particulares ocupando cada tipo de habitat. Neste estudo nós apresentamos as mudanças na composição, riqueza de espécies e abundância de aves através de um gradiente de florestas, cerrados e campos de altitude no Maciço do Urucum, uma região montanhosa localizada no domínio das Florestas Secas Chiquitanas no oeste do Brasil. Nós registramos 165 espécies de aves por métodos quantitativos e qualitativos. Savanas florestadas, florestas ripárias e florestas submontanas apresentaram a maior riqueza e abundância de aves, enquanto savanas arborizadas e campos de altitude tiveram valores intermediários e baixos, respectivamente. A composição de espécies foi similar entre florestas ripárias e submontanas, enquanto os outros tipos de vegetação apresentaram comunidades de aves mais dissimilares. Nossos resultados demonstraram diferenças na composição, riqueza e abundância de aves entre os tipos de vegetação presentes no Maciço do Urucum, e apontaram a importante função dos gradientes de vegetação para a conservação das comunidades de aves em montanhas. Adicionalmente, este é o primeiro estudo sobre as comunidades de aves nas Florestas Secas Chiquitanas brasileiras, um importante domínio que ocorre no oeste do Brasil e que tem sido pouco estudado.

ABSTRACT Different vegetation types are distributed in mountains according to altitude, topography and soil. The composition and structure of bird communities in these areas can change in relation to the vegetation gradient, with particular communities occupying each habitat type. In this study we present the changes in composition, species richness and bird abundance over the gradient of forests, savannas and altitudinal grasslands of Maciço do Urucum, a mountainous region located in the Chiquitano Dry Forests domain in western Brazil. We recorded 165 bird species through qualitative and quantitative methods. Forested savannas, riparian forests and submontane forests presented the highest richness and abundance of birds, while arboreal savannas and altitudinal grasslands had intermediate and low values, respectively. The bird composition was similar between riparian and submontane forests, while other vegetation types present more dissimilar bird communities. Our results show differences in composition, richness and bird abundance among the vegetation types present at Maciço do Urucum, and highlight an important function of vegetation gradients for the conservation of bird communities in mountains. Additionally, this is the first study of the bird communities in the Brazilian Chiquitano Dry Forests, an important domain in the west of Brazil which has been poorly studied.