@InProceedings{CastroBezeVonRVonR:2023:GrPrPr,
               author = "Castro, Aline Anderson de and Bezerra, Francisco Gilney Silva and 
                         Von Randow, Rita and Von Randow, Celso",
          affiliation = "{Instituto Nacional de Pesquisas Espaciais (INPE)} and {Instituto 
                         Nacional de Pesquisas Espaciais (INPE)} and {Instituto Nacional de 
                         Pesquisas Espaciais (INPE)} and {Instituto Nacional de Pesquisas 
                         Espaciais (INPE)}",
                title = "Gross Primary Productivity heterogeneity in the Amazon forest and 
                         possible vulnerabilities to climate change",
            booktitle = "Proceedings...",
                 year = "2023",
         organization = "AGU FAll Meeting",
            publisher = "AGU",
             abstract = "Brazil concentrates in the Amazon biome its main emissions 
                         associated with deforestation and forest degradation, but also 
                         possible sinks through the natural forest. This study focuses on 
                         understanding this sink role and the factors that determine 
                         resilience or vulnerability of the biome to future climate change. 
                         To achieve this, we used Gross Primary Productivity (GPP) obtained 
                         by the GOSIF Gross Primary Productivity based on OCO-SIF, MODIS 
                         remote sensing data as a forest productivity proxy. We applied 
                         k-means clustering method to classify regions according to the 
                         annual mean GPP and also to its amplitude (difference between 
                         maximum and minimum monthly mean) in natural forest. For each 
                         cluster, we evaluated the main atmospheric data that can explain 
                         the GPP using linear regression over a set of variables obtained 
                         from ERA5-Land reanalysis data. To remove deforested areas, we 
                         classified the 2020 Pan-Amazonian land cover Mapbiomas data as 
                         natural or non-natural areas to produce a mask of only natural 
                         areas. The main conditions found as GPP explanatory variables, as 
                         expected, are evaporation, solar radiation, precipitation, and 
                         soil related, such as heat flux, soil water content and 
                         temperature and albedo. Either clustering by the annual mean or by 
                         the annual amplitude, the central Amazon appears mainly influenced 
                         by forest evaporation, while the southern areas are more 
                         influenced by the soil properties. Precipitation is a relevant 
                         driver in the southwestern and eastern areas, together with solar 
                         radiation. The main difference between the two clustering methods 
                         was found in the northwestern areas, where the clustering based on 
                         mean GPP is related to evaporation and the clustering based on GPP 
                         amplitude shows a stronger influence of the radiation and, to a 
                         lesser extent, surface runoff and relative humidity. The next step 
                         is to include other ecological or physiological factors in the GPP 
                         distribution over the Amazon Basin, to find out if there are any 
                         particularities in the main drivers (meteorological and/or 
                         ecological) of GPP variability in different areas of the amazon.",
  conference-location = "San Francisco, CA",
      conference-year = "11-15 Dec. 2023",
             language = "en",
        urlaccessdate = "28 abr. 2024"
}