Supplement to GSWP-2: Details of the Forcing Data
Supplement to GSWP-2: Details of the Forcing Data
复制标题
GSWP-2 的补充：强迫数据的详细信息
DOI：
10.1175/bams-87-10-dirmeyer
复制
发表时间：
2006
影响因子：
8
通讯作者：
N. Hanasaki
中科院分区：
文献类型：
--
作者：
P. Dirmeyer;Xiang Gao;Mei Zhao;Zhichang Guo;T. Oki;N. Hanasaki
表S1-S3给出了LSS输入字段的完整列表。本文介绍了表S3所列气象强迫资料的推导过程。赵和Dirmeyer(2003)对强迫数据给出了完整的描述。在这里，我们给出一个总结，让读者熟悉我们使用的观测数据和模型(再分析)数据的混合过程，并指出最终可能影响第二次全球土壤湿度项目(GSWP-2)分析的问题。所有GSWP-2模型都在国际卫星陆地表面气候学项目(ISLSCP)倡议II 1网格的陆地网格点上运行。来自国家环境预测中心(NCEP)/能源部(DOE)再分析的3小时近地表气象数据(Kanamitsu等人)。为了在ISLSCP 1网格上共同登记T62(高斯网格)分辨率(全球192×94个网格框)NCEP/DOE再分析，再分析数据集与每个网格的陆地-海洋掩模定义尽可能一致地重新加密。NCEP/DOE陆海掩码和ISLSCP陆海掩码用于确保陆点转换为陆点，海点转换为海点。对于目标ISLSCP栅格上的每个栅格框，输入NCEP/DOE栅格有一个、两个或四个重叠的栅格框；只选择与目标栅格点类型相同的点(海代表海洋，陆地代表陆地)。所有相同类型的相交网格框用于执行数据的双线性内插。如果所有相交的栅格框都不具有相同的类型，则它们都用于双线性插补。这可能发生在例如ISLSCP掩模有一个湖，而NCEP/DOE掩模在周围只有陆点的位置。在这种情况下，陆地和海洋面具之间的一致性就会丢失。图S1显示了两个掩码之间的一致性。在ISLSCP计划II有土地但重叠的NCEP/DOE网格只有水的地方，网格框是红色的，绿色阴影表示ISLSCP计划II有水但NCEP/DOE有陆地的地方。全球观测数据集，如降水量，可能优于对降水量的再分析估计，但它们远非完美。仪器覆盖不完整、难以收集观测数据、各国在仪器和校准方面存在差异以及仪器本身存在问题，这些都可能导致观测数据的覆盖范围和质量非常不均衡。Oki等人(1999)表明，在GSWP-1中，系统地低估了高纬度地区的径流，因为那里的降雪量对年降雨量有很大贡献。Motoya等人(2002)对此进行了研究
Acomplete list of the input fields for the LSSs is given in Tables S1–S3. The derivation of the meteorological forcing data listed in Table S3 is described here. A complete description of the forcing data is given in Zhao and Dirmeyer (2003). Here we present a summary to acquaint the reader with the process of hybridization of observational and model (reanalysis) data that we have used, and to point out issues that may ultimately bear on the Second Global Soil Wetness Project (GSWP-2) analysis. All GSWP-2 models were run on the land grid points of the International Satellite Land Surface Climatology Project (ISLSCP) Initiative II 1 grid. The 3-hourly near-surface meteorological data from National Centers for Environmental Prediction (NCEP)/Department of Energy (DOE) reanalysis (Kanamitsu et al. 2002) was regridded for ISLSCP Initiative II. To coregister the T62 (Gaussian grid) resolution (192× 94 grid boxes globally) NCEP/DOE reanalysis on the ISLSCP 1 grid, the reanalysis dataset was regridded as consistently as possible with the land–sea mask definitions of each grid. The NCEP/DOE land–sea mask and the ISLSCP land–sea mask are used to ensure that land points are transformed into land points and sea points are used for sea points. For every grid box on the target ISLSCP grid, there are either one, two, or four overlapping grid boxes of the input NCEP/DOE grid; only those points are selected that are of the same type as the target grid point (sea for sea and land for land). All intersecting grid boxes of the same type are used to perform a bilinear interpolation of the data. If none of the intersecting grid boxes have the same type, they are all used for bilinear interpolation. This may occur at locations where, for instance, the ISLSCP mask has a lake, whereas the NCEP/DOE mask has only land points in the surroundings. In this case, the consistency between land–sea masks is lost. Figure S1 shows the consistency between the two masks. Grid boxes are shaded red where ISLSCP Initiative II has land but the overlapping NCEP/DOE grid has only water, and green shading indicates where ISLSCP Initiative II has water but NCEP/DOE has land.Precipitation. Global observational datasets of quantities such as precipitation are probably superior to reanalysis estimates of precipitation, but they are far from perfect. Incomplete gauge coverage, difficulty in collecting observational data, national variations in instruments and calibration, and problems with the instruments themselves can lead to very uneven coverage and quality of observational data. Oki et al.(1999) showed that in GSWP-1, runoff was systematically underestimated over high latitudes where snow is a significant contributor to annual precipitation. Motoya et al.(2002) have examined this