A surface wind speed map for Mexico based on NARR and observational data

Abstract

A systematic comparison of the surface wind speed records extracted from the North American Regional Reanalysis (NARR) database and the network of automated surface observation stations of the Mexican National Weather Service has been conducted. The work presented here is based on a multi-step procedure that initiates with a point-to-point comparison of surface observations and NARR grid point, followed by the reconstruction of missing data from the quality-controlled observational data set using an iterative technique based on principal component analysis (PCA). Subsequently, the empirical orthogonal function (EOF) sets for NARR and the reconstructed observational data set were compared globally using a group comparison criterion that is also based on PCA. Different observational periods were evaluated, leading to the highest similarity between the two sets for the case in the year 2010. A wind speed map for Mexico was created by joining the quality-controlled, filtered and reconstructed observational data with the NARR map using a weighting technique based on distance weighting, local similarity of the EOFs and data recovery.

Comparison of ERA5 surface wind speed climatologies over Europe with observations from the HadISD dataset

“…These results are in accordance with other studies (Decker et al ., 2012; Cannon et al ., 2015; Kaiser‐Weiss et al ., 2015; Rose and Apt, 2016; Minola et al ., 2020), where wind speeds from different reanalyses products are compared against flux tower or meteorological data in different areas of the globe. But in contrast with others (Toledo et al ., 2015; Coburn, 2019), probably mainly related to the studied area. Coburn (2019) in the upper Midwest of the United States, showed that correlations between reanalyses and observations improve with higher temporal reanalysis resolution (monthly correlations are smaller than daily correlations).…”

Towards Better Wind Resource Modeling in Complex Terrain: A k-Nearest Neighbors Approach

“…The transfer of wind roses may not only change the Weibull parameters, but may also distort the underlying histogram, resulting in varying degrees of goodness of fit. R 2 -related parameters can be constructed for their use as similarity features; additionally, a modified weighting scheme (similarly to the one proposed in [19]) was explored as well, where the R 2 -value was used to build a confidence level matrix as part of a generalized inverse-distance averaging scheme.…”