If rain falls in the forest and no one is around to record that it fell, do the data exist? If this rain falls anywhere from the far southern tip of Mexico to the extreme northern reaches of the Canadian Arctic (or in Puerto Rico and Hawaii), these data are part of a collection called Daymet, and are only a mouse-click away.
Accessible through NASA’s Oak Ridge National Laboratory Distributed Active Archive Center (ORNL DAAC), Daymet provides near-surface meteorological information in remote areas or in areas with limited instrumentation. Daymet Version 2 data were first made available through the ORNL DAAC in 2013, and meteorological data are available dating back to 1980. “Daymet was created to support research on some specific types of ecological models,” says Dr. Bruce Wilson, ORNL DAAC manager. “However, the usage has gone way beyond that original niche.”
Prior to the development of Daymet, meteorological data were acquired from individual weather stations. These point data were then extrapolated to estimate meteorological conditions in study areas that did not have weather stations. This technique was relatively straightforward for variables like temperature and precipitation, but required advanced relationships for more complex variables like humidity and radiation. The Daymet algorithm incorporates a varying search radius that enables the integration of data from multiple weather stations for each predicted grid cell. “Daymet filled the need to have a gridded, continuous dataset of near-surface meteorological conditions to provide inputs for driving terrestrial ecosystem models,” explains ONRL DAAC Daymet lead Michele Thornton.
Seven weather parameters are available through Daymet: daily high and low temperature, precipitation, vapor pressure, shortwave radiation, snow water equivalent, and day length. These daily data are produced on a 1 km x 1 km gridded surface, which provides an exceptionally high level of detail. With data available for every square kilometer of the continent, this makes for a dataset of approximately 500 million data points for each year in the series.
Daymet remains one of the most popular data collections at the ORNL DAAC, which is a partnership between NASA and the U.S. Department of Energy (DOE) and responsible for archiving and distributing data in NASA’s Earth Observing System Data and Information System (EOSDIS) collection related to biogeochemical dynamics, ecology, and environmental processes. According to metrics from NASA’s Earth Science Data and Information System (ESDIS) Project, more than 112 terabytes of Daymet data were distributed during the 2019 Fiscal Year (October 1, 2018 through September 30, 2019).
Ongoing development of Daymet, its archiving, and its distribution are supported by the ESDIS Project and Terrestrial Ecology Program. Daymet algorithm development and processing also are supported by the Biological and Environmental Research program within the DOE’s Office of Science.
Daily meteorological data for 2019 recently were incorporated into the current Daymet Version 3. “We produce annual updates of Daymet after a full calendar year,” says Thornton. “I wait until weather station inputs are of ‘archive quality’ (about mid-February) and it takes a few weeks to process all the data and standardize the file formats for our various access methods. Researchers are always anxious to have these data.”
The initial response to the addition of 2019 data validates Thornton’s point. In the span of just one week, from March 22 to March 28, 2020, more than one million individual queries for 2019 data were made using Daymet’s Single Pixel Extraction Tool, according to ORNL DAAC metrics. The Single Pixel Extraction Tool enables users to acquire daily data from the nearest 1 km x 1 km Daymet grid cell for a single geographic point by latitude and longitude in decimal degrees, and is one of several tools for using Daymet data at the ORNL DAAC. The Daymet data delivery tools also include the option to download comprehensive cross-validation statistics, so users can evaluate the uncertainty associated with Daymet data and decide if the quality is acceptable for use in their particular application. “Many of our users ask if we can update the data more often,” says Dr. Wilson. “We are working on this, and we recognize that there are many applications where more up-to-date Daymet data will provide even more value to our users.”
The value of Daymet data is evident in the many investigations and applications in which Daymet data are being used. “Daymet data are used in a wide variety of studies: hydrology, vegetation and crop modelling, wildlife research, even pathogen distribution,” Thornton says.
Almost 325 peer-reviewed papers have used some aspect of Daymet data since 2012, according to ORNL DAAC records as of April 27, 2020. An ORNL DAAC-written article published on the NASA Earthdata website spotlights some recent uses of Daymet data. These include peer-reviewed research into the potential effects of climate change on common milkweed plants and responses of vegetation to climate variability across Yellowstone National Park along with the incorporation of Daymet data into USGS hydrologic modeling applications. The popularity of Daymet also has led to the creation of community-developed open source scientific software such as daymetr (an R package) and daymetpy (a Python package), both of which are available through the ORNL DAAC Daymet Resources Learning page.
Daymet development continues, and generation of Daymet Version 4 data products already is underway with a target release date at the ORNL DAAC of Fall 2020. As Thornton notes, Daymet Version 4 enhancements will help improve the overall accuracy and precision of the Daymet continuous surfaces, and will include attention to weather station time-of-observation bias, improved constraints on horizontal interpolation coefficients, and modification of core model algorithms to improve interpolation in regions with low station density and during conditions of extreme precipitation, such as hurricanes.
“There are so many studies that benefit from knowledge of historic daily weather drivers and the influence of these drivers on the study area or the species being studied,” says Thornton. “Having access to daily data also means that researchers can derive any multitude of climatic variables based on their time frame of interest.”
Thanks to Daymet, if precipitation falls, snow accumulates, or radiation inputs change in North America—from Tapachula, Mexico, to Alert, Canada, or anywhere in-between—these data will not only be collected, but will be added to a nearly 40-year data collection enabling a wide range of ecological studies.
Explore Daymet data at ORNL DAAC: https://daac.ornl.gov/daymet
Learn how to use Daymet and Daymet data analysis tools: https://daymet.ornl.gov/learning