The Amazon region of Brazil is a vitally important region for water and carbon cycling both for the region and the globe. This region is experiencing the impacts of global climate change as well as local land cover changes. Here, we investigated water and carbon estimates and related remotely sensed variables from both MODIS satellite and the Noah-MP land surface model for 3 years (2015–2017) in the state of Mato Grosso, Brazil. Land surface temperature agrees well between MODIS and the model, while the leaf area index (LAI) is higher in the model simulations. The monthly evapotranspiration (ET) from MODIS (MOD16A2) and gross primary productivity (GPP, MOD17A2) were lower than, but well correlated with, the model simulations. A noticeable exception was in the Broadleaf Forest class, which accounts for approximately 50% of the land cover in the state, where the modeled LAI was out of phase with the satellite observations, resulting in significantly poorer performance in the water and carbon fluxes for that land cover class. In addition, we investigated the sensitivity of the ET and GPP to precipitation forcing. The modeled ET relationships show correlations of approximately 0.6 for all classes (Broadleaf Forest being the exception, 0.24), while the MODIS shows reduced values averaging about 0.5 (Broadleaf Forest = 0.03). The slopes of the relationships illustrated the same sensitivity between MODIS and Noah-MP with the exception of Grasslands and Open Shrublands. The GPP relationships with precipitation show lower correlations across all land cover types for both MODIS and Noah-MP, with the slopes being significantly different for the Open Shrublands and Broadleaf Forest classes. In each of these classes, the Noah-MP simulations resulted in increased sensitivity to precipitation than was observed in the MODIS products. We highlight that this comparison is essential for increasing our understanding of how these different sources estimate water and carbon cycling and can be utilized for assessing the impacts of climate and land cover change in the region.

巴西的亚马逊地区是该地区和全球水和碳循环至关重要的地区。该地区正在遭受全球气候变化以及当地土地覆盖变化的影响。在这里，我们调查了巴西马托格罗索州（MODIS）卫星和Noah-MP地表模型中水和碳的估算值以及相关的遥感变量，历时3年（2015-2017年）。陆地表面温度在MODIS和模型之间非常吻合，而在模型仿真中叶面积指数（LAI）较高。MODIS（MOD16A2）的月蒸散量（ET）和总初级生产力（GPP，MOD17A2）低于模型仿真，但与模型仿真相关性很高。一个明显的例外是阔叶林类，该模型约占该州土地覆盖的50％，因为模拟的LAI与卫星观测值不一致，导致该土地覆盖类别的水和碳通量性能显着下降。此外，我们研究了ET和GPP对降水强迫的敏感性。建模的ET关系显示所有类别之间的相关性约为0.6（Broadleaf Forest是例外，0.24），而MODIS显示的平均降低值约为0.5（Broadleaf Forest = 0.03）。这些关系的斜率说明了MODIS和Noah-MP之间的敏感性相同，除了草地和开放灌木地。GPP与降水的关系表明，对于MODIS和Noah-MP，所有土地覆盖类型的相关性均较低，对于开放灌木丛和阔叶林类别，坡度明显不同。在这些类别的每一个类别中，与MODIS产品相比，Noah-MP模拟都增加了对降水的敏感性。我们着重指出，这种比较对于增进我们对这些不同来源如何估算水和碳循环的理解至关重要，并可用于评估该地区气候和土地覆盖变化的影响。