Climate Impacts of Land Use and Land Cover Using Subgrid Information from Earth System Models

美国地表气温地图，意味着在地表以上2米处记录的气温，通常有一个粗略的外观，南部的气温更高。 但是，表面本身的温度可以在短短几步之内发生巨大变化，比如赤脚从草地野餐区穿过炎热的沙滩到海滩的冷水。 这种不匹配的空间尺度之间的不均匀的土地表面和广阔的大气层之间的理解是一个挑战，两者如何相互作用：每一片土地与它上面的空气片直接相互作用，但大气以某种方式平滑斑块和响应的总效应的水分，热量，和其他输入从表面。 作为一个实际问题，这种聚合的一些必须明确编码到天气和气候模式，因为它是不可能的，代表每一个独特的补丁的当地陆地表面上的全球网格的天气或气候模式。 这种聚合的标准策略是通过在每个表面网格正方形（网格正方形可以是50公里宽或更宽）内使用多个“瓦片”来表示表面斑块。瓦片代表不同的地表类型，例如铺砌的城市环境、农场、森林和湖泊，每种都具有适当的反射率、粗糙度、排水量和调节陆地-大气耦合的其他参数值。 每个瓦片覆盖的网格正方形的百分比被指定，并且热量、水分、动量和成分的交换在瓦片分数上被求和，以形成与大气共享的网格总数。 气候模拟产生了大量的瓦片数据，这可以为了解陆-气相互作用及其气候后果提供宝贵的资源。 然而，瓦片数据很少与气候模式输出一起存档，通常只包含网格平方和。本项目旨在利用瓦片级数据的未开发资源，了解陆-气耦合及其对大尺度气候的影响。 作为土地利用模型相互比较项目的一部分，PI组织了收集这些数据的工作，以便可以对多个模型的结果进行比较。 数据收集两个具体的模拟，一个考虑森林砍伐的影响，假设一个固定的二氧化碳（CO2）浓度，另一个在CO2大幅增加后的未来排放量（SSP 3）的情况下，在研究中的一个关键问题是在气候的差异，可以与土壤水库的水分，热量和养分的代表性。 在常用的“共享列”方法中，网格框内的所有瓦片共享相同的土壤储层，而“独立列”方法对网格框内的每个瓦片使用单独的土壤储层。 PI认为，独立列近似导致更极端的网格箱表面温度，例如，树木覆盖的瓷砖只能获得自己瓷砖内的土壤水分，因此更容易发生干旱。 尽管如此，这可能是一个更好的水文表现，因为具有相似功能特性的植物经常被发现在一起，水共享的程度可能是有限的，因为网格框通常比水可以渗透和热量可以扩散的距离大得多。 它是研究陆-气耦合及其在气候中作用的研究人员社区的一个资源，也是寻求改进用于为决策者提供天气预报和气候预测的模型的模型开发人员的一个资源。 瓦片数据还可用于评估气候变化对具有特定地面覆盖的区域可能产生的气候后果，例如，数据可揭示气候变暖时草原和森林之间水资源可用性的特征差异。 该项目还支持一名研究生和一名博士后，从而为该研究领域的未来劳动力提供支持。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

A map of US surface air temperature, meaning the air temperature recorded 2 meters above the surface, typically has a broad-brush appearance with warmer temperatures to the south. But the temperature of the surface itself can vary dramatically in a few short steps, say on the barefoot walk from a grassy picnic area across hot sand to the cool water at the beach. This mismatch in spatial scale between the patchy land surface and the broad-brush atmosphere poses a challenge for understanding how the two interact: each patch of land interacts with the patch of air directly above it, yet the atmosphere somehow smooths out the patchiness and responds to the aggregate effects of the moisture, heat, and other inputs it receives from the surface. As a practical matter some of this aggregation must be explicitly coded into weather and climate models as it is not possible to represent every distinct patch of the local land surface on the global grid of a weather or climate model. The standard strategy for such aggregation is to represent surface patchiness through the use of multiple "tiles" within each surface grid square (the grid squares can be 50 kilometers wide or wider). Tiles represent distinct land surface types, for instance paved urban environments, farms, forests, and lakes, each with appropriate values of reflectivity, roughness, drainage, and other parameters that regulate land-atmosphere coupling. The percentage of the grid square covered with each tile is specified, and the exchanges of heat, moisture, momentum, and constituents are summed over the tile fractions to form grid totals which are shared with the atmosphere. Climate simulations generate a tremendous amount of tile data, which could provide a valuable resource for understanding land-atmosphere interactions and their climatic consequences. Yet tile data is rarely archived with climate model output, which generally contains only the grid square totals.This project seeks to exploit the untapped resource of tile-level data for understanding land-atmosphere coupling and its implications for large-scale climate. The PIs have organized an effort to collect this data as part of the Land Use Model Intercomparison Project (LUMIP) so that results from multiple models can be compared. Data is collected for two specific simulations, one which considers the effects of deforestation assuming a fixed carbon dioxide (CO2) concentration, and another in which CO2 increases substantially following a scenario of future emissions (SSP3).A key issue in the research is differences in climate that can be related to the representation of soil reservoirs of moisture, heat, and nutrients. In the commonly used "shared column" approach all tiles within a grid box share the same soil reservoir, while the "independent column" approach uses a separate soil reservoir for each tile within a grid box. The PIs argue that the independent column approximation leads to more extreme grid box surface temperature as, for example, a tree-covered tile only has access to soil moisture within its own tile and is thus more prone to drought. Nevertheless this may be a better representation of the hydrology as plants with similar functional traits are often found together, and the extent of water sharing across a grid box may be limited as grid boxes are typically much larger than the distances over which water can percolate and heat can diffuse.The database of tile outputs from multiple models constitutes an important broader impact of the project. It serves as a resource for the community of researchers working on land-atmosphere coupling and its role in climate, and for model developers seeking to improve models used to provide weather predictions and climate projections for decision makers. The tile data can also be used to assess the likely climatic consequences of climate change over regions with particular ground cover, for instance the data can reveal characteristic differences in water availability between grasslands and forests in a warming climate. The project also supports a graduate student and a postdoc, thereby providing for the future workforce in this research area.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.