Daily 2 P.M. Average Relative Humidity Time Series Maps and Point Data for State of Hawaii

Description

This dataset contains time series maps of daily 2 p.m. average terrestrial relative humidity for four counties of the State of Hawaii spanning the time period from Jan. 1. 2002 through December 31, 2025 with a uniform spatial resolution of 250 m. x 250 m. Each map in the data_maps folder represents one day of 2 p.m. gridded relative humidity (representing the averaged relative humidity from 1:00 p.m. through 1:59 p.m.), provided in GeoTiff raster format containing one data layer and a corresponding land-sea mask layer. Point data time series of the 2 p.m. averaged relative humidity for all weather stations which reported data on each day of the time series associated with the maps are included as column-separated value (CSV) tables, as well as CSVs of corresponding leave-one-out cross-validation metrics for the maps with respect to the observed data. The maps are derived by using a multi-adaptive regression spline model to spatially interpolate the observed weather station point data to the map grid. These data are intended to support any research or applications that require spatially-continuous, high resolution relative humidity data in the State of Hawaii.

Steps to reproduce

The methods and a full data description are intended to be published in a data article in the Data In Brief. The article abstract is shown below. This dataset will be updated with an article doi pending review and journal acceptance. Abstract: Gridded daily relative humidity (RH) data are produced as a 2 p.m. average for the Hawaiian Islands from 2002–2025 (24 years) with a resolution of 250 m. A multi-adaptive regression spline (MARS) model is used to spatially interpolate daily RH from weather station observations. The gridded data are produced daily by fitting the MARS model to the 2 p.m. averages of the dew point temperatures measured each day, using their elevation as the sole predictor for the model. The piece-wise linear model derived from this process is then applied to a statewide digital elevation model (DEM) to estimate the dew point values at the DEM grid points. The resulting maps of dew point temperature were converted to RH using concurrent air temperature estimated via linear regression from pre-existing gridded daily maximum air temperature time series. A dataset of quality-controlled 2 p.m. RH station observations, which were used to estimate each corresponding map have been provided with the gridded data. This dataset was developed primarily for wildfire risk monitoring but additionally has applications for estimating evapotranspiration and soil moisture, each of which have significant value for drought monitoring and water resource management across the state.

Institutions

• University of Hawaii System

  Hawaii, Honolulu

Categories

Funders

• U.S. National Science Foundation

  Government of the United States of America

  Virginia, Alexandria

  Grant ID: OIA-2149133

Licence