Abstract

The exponential growth in solar radiation measuring stations across the conterminous United States permits the generation of gridded solar irradiance data that capture the spatiotemporal variability of solar irradiance far more accurately than was previously possible from ground-based observations. Taking advantage of these observations, we generated a 30-yr climatology (1991–2020) of mean monthly global irradiance at a resolution of 30 arc s (∼800 m) on both a horizontal surface and a sloped ground surface. This paper describes the methods used to generate the gridded data, which include extensive quality control of station data, spatial interpolation of effective cloud transmittance using the “PRISM” method, and simulation of the effects of elevation, shading, and reflection from nearby terrain on solar irradiance. A comparison of the new dataset with several other solar radiation products reveals some spatial features in solar radiation that are either lacking or underresolved in some or all of the other datasets. Examples of these features include strong gradients near foggy coastlines and along mountain ranges where there is persistent orographically driven cloud formation. The workflow developed to create the long-term means will be used as a template for generating time series of monthly and daily solar radiation grids up to the present.

The Cooperative Observer Program (COOP), established over 100 years ago, has become the backbone of temperature and precipitation data that characterize means, trends, and extremes in U.S. climate. However, significant and widespread biases in the way COOP observers measure daily precipitation have been discovered. These include 1) underreporting of light precipitation events (daily totals of less than 0.05 in., or 1.27 mm), and 2) overreporting of daily precipitation amounts evenly divisible by five- and/or ten-hundredths of an inch, that is, 0.10, 0.25, 0.30 in., etc. (2.54, 6.35, 7.62 mm, etc.). Observer biases were found to be highly variable in space and time, which has serious implications for the spatial and temporal trends and variations of commonly used precipitation statistics. In addition, it was found that few COOP stations had sufficiently complete data to allow the calculation of stable precipitation statistics for a stochastic weather simulation model. Out of more than 12,000 COOP stations nationally, only 784 (6%) passed data completeness and observer bias screening tests for the climatological period 1971–2000. Of the 1221 COOP stations selected for the U.S. Historical Climate Network (USHCN), which provides much of the country's official data on climate trends and variability over the past century, only 221 stations (18%) passed these tests. More effective training materials and regular communication with COOP observers could reduce observer bias in the future. However, it is unlikely that observer bias can be eliminated. One solution is to automate the COOP precipitation measurement system, but this is an expensive option, and may increase other biases associated with automated precipitation measurement. Further analyses are needed to better quantify and characterize observer bias, and to develop methods for dealing with its effects.

Abstract

There is a great need for gridded daily precipitation datasets to support a wide variety of disciplines in science and industry. Production of such datasets faces many challenges, from station data ingest to gridded dataset distribution. The quality of the dataset is directly related to its information content, and each step in the production process provides an opportunity to maximize that content. The first opportunity is maximizing station density from a variety of sources and assuring high quality through intensive screening, including manual review. To accommodate varying data latency times, the Parameter-Elevation Regressions on Independent Slopes Model (PRISM) Climate Group releases eight versions of a day’s precipitation grid, from 24 h after day’s end to 6 months of elapsed time. The second opportunity is to distribute the station data to a grid using methods that add information and minimize the smoothing effect of interpolation. We use two competing methods, one that utilizes the information in long-term precipitation climatologies, and the other using weather radar return patterns. Last, maintaining consistency among different time scales (monthly vs daily) affords the opportunity to exploit information available at each scale. Maintaining temporal consistency over longer time scales is at cross purposes with maximizing information content. We therefore produce two datasets, one that maximizes data sources and a second that includes only networks with long-term stations and no radar (a short-term data source). Further work is under way to improve station metadata, refine interpolation methods by producing climatologies targeted to specific storm conditions, and employ higher-resolution radar products.