Abstract. Every year, a dense smoke haze covers a large portion of South America originating from fires in the Amazon Basin and central parts of Brazil during the dry biomass burning season between August and October. Over a large portion of South America, the average aerosol optical depth at 550 nm exceeds 1.0 during the fire season, while the background value during the rainy season is below 0.2. Biomass burning aerosol particles increase scattering and absorption of the incident solar radiation. The regional-scale aerosol layer reduces the amount of solar energy reaching the surface, cools the near-surface air, and increases the diffuse radiation fraction over a large disturbed area of the Amazon rainforest. These factors affect the energy and CO 2 fluxes at the surface. In this work, we applied a fully integrated atmospheric model to assess the impact of biomass burning aerosols in CO 2 fluxes in the Amazon region during 2010. We address the effects of the attenuation of global solar radiation and the enhancement of the diffuse solar radiation flux inside the vegetation canopy. Our results indicate that biomass burning aerosols led to increases of about 27 % in the gross primary productivity of Amazonia and 10 % in plant respiration as well as a decline in soil respiration of 3 %. Consequently, in our model Amazonia became a net carbon sink; net ecosystem exchange during September 2010 dropped from +101 to −104 TgC when the aerosol effects are considered, mainly due to the aerosol diffuse radiation effect. For the forest biome, our results point to a dominance of the diffuse radiation effect on CO 2 fluxes, reaching a balancePublished by Copernicus Publications on behalf of the European Geosciences Union. 14786 D. S. Moreira et al.: Modeling the radiative effects of biomass burning aerosols of 50-50 % between the diffuse and direct aerosol effects for high aerosol loads. For C3 grasses and savanna (cerrado), as expected, the contribution of the diffuse radiation effect is much lower, tending to zero with the increase in aerosol load. Taking all biomes together, our model shows the Amazon during the dry season, in the presence of high biomass burning aerosol loads, changing from being a source to being a sink of CO 2 to the atmosphere.

Processos estocásticos de natureza espaço-temporais consistem de fenômenos que são caracterizados por meio da variabilidade espacial e temporal. Atualmente, é uma das áreas de maior crescimento com diversas aplicações em ciências ambientais, geográficas, biológicas, epidemiológicas, entre outras. Certamente, os métodos da estatística convencional não são adequados para modelar estruturas autocorrelacionadas no espaço e no tempo. De fato, ainda há grandes desafios no tange à implementação computacional da metodologia geoestatística para análise de processos espaços-temporais, com destaque para o pacote spacetime do programa R, utilizado neste estudo. Assim, este trabalho tem como objetivo aplicar a metodologia geoestatística espaço-temporal de funções de covariância a fim de inferir acerca da temperatura máxima do ar do Estado de Minas Gerais de 1996 a 2016, visando contribuir com desafios, tais como aquecimento global, urbanização descontrolada, escassez de recursos naturais, epidemias e catástrofes naturais. Utilizando os dados de 61 estações meteorológicas foi realizada a análise geoestatística espaço-temporal, no qual o modelo de covariância somamétrico foi o mais adequado, considerando-se o critério do Erro Quadrático Médio. Dessa forma, foi possível elaborar mapas de predições das temperturas máximas do ar no estado de Minas Gerais por meio da krigagem ordinária, assumindo-se estacionariedade de primeira ordem do processo estocástico avaliado. Pode-se observar que os modelos da geoestatística espaço-temporal mostraram ser eficientes nos estudos espaço-temporais das temperaturas máximas do ar. Palavras-chave: Modelagem de Dados Espaço-Temporal, Covariância, variograma, Krigagem Ordinária.

<p><strong>Abstract.</strong> Every year, a dense smoke haze of regional dimensions covers a large portion of South America originated from fire activities in the Amazon Basin and Central parts of Brazil during the dry/biomass-burning season between August and October. Over a large portion of South America, the average aerosol optical depth at 550 nm exceeds 1.0 during the fire season while the background value during the rainy season is below 0.2. Smoke aerosol particles increase scattering and absorption of the incident solar radiation. The regional-scale aerosol layer reduces the amount of solar energy reaching the surface, cools the near surface air, and increases the diffuse radiation fraction over a large disturbed area of the Amazon rainforest. These factors affect the energy and CO<sub>2</sub> fluxes at the surface. In this work, we applied a fully integrated atmospheric model to assess the impact of smoke aerosols in CO<sub>2</sub> fluxes in the Amazon region during 2010. We address the effects of the attenuation of the solar global radiation and the enhancement of the diffuse solar radiation flux inside the canopy. Our results indicated that the smoke aerosols led to an increase of about 22 % of the gross primary productivity of Amazonia, 9 % of plant respiration and a decline in soil respiration from of 3 %. Consequently, Amazonia net ecosystem exchange during September 2010 dropped from +101 to −104 TgC when the aerosol effects were considered, mainly due to the aerosol diffuse radiation effect. For the forest biome, our results pointed to a dominance of the diffuse radiation effect on CO<sub>2</sub> fluxes, reaching a balance of 50 % – 50 % between the diffuse and direct aerosol effects for high aerosol loads. For C3 grass type and <i>cerrado</i>, as expected, the contribution of the diffuse radiation effect is much lower, tending to zero with the increase of aerosol load. That is, the Amazon during the dry season, in the presence of high smoke aerosol loads, change from being a source to be a sink of CO<sub>2</sub> to the atmosphere.</p>