Datasets
The foundation of the seasonal statistical forecast models is the dataset used to train the model. PyForecast uses the approach of examining a large dataset to determine the best predictive model without prescreening, which has been shown to result in artifical skill (DelSole and Shukla, 2009). The datasets below are available for retrieval through PyForecast dataloaders and may be useful for forecast development.
Most SWE measurement data can be found at the U.S. Department of Agriculture-Natural Resources Conservation Service's (NRCS) interactive map found here.
Automated and manual SWE measurements are taken throughout the western United States. The following sections describe these measurements.
The SNOTEL network is maintained by the U.S. Department of Agriculture-Natural Resources Conservation Service (NRCS). SNOTEL monitors snowpack, precipitation, air temperature, and soil moisture in mountainous areas. These data are frequently used in statistically based forecasts.
The primary measurement from SNOTEL sites used for forecasting snowmelt-dominated basins is snow water equivalent (SWE). SWE measures the water content of the snowpack.
Automated Snow Monitoring-Natural Resources Conservation Service
British Columbia also collects automated snow measurements, which can be found here.
NRCS also measures snowpack manually throughout the western United States at a number of locations. Snow course measurements comprise a series of snowpack samples at a recurring interval, typically on or around the first of the month. More information regarding snow courses can be found at:
Manual Snow Monitoring: Natural Resources Conservation Service
Some state agencies also perform snow surveys, including:
SNODAS is a gridded product which blends remotely sensed and modeled snowpack data, as described in (Carroll et al., 2006). PyForecast currently does not have the ability to incorporate SNODAS data in its forecasts. The State of Colorado's CDSS processes SNODAS grids in a manner suitable for PyForecast ingestion.
Both gridded and point measurements of precipitation and air temperature may prove useful for forecasting streamflow. According to (NRCS, 2011), air temperature may indicate the acceleration or delay of snowmelt, and precipitation is directly related to runoff. Typically, these variables are accumulated over a specified timeframe. For example, a forecaster may use mean winter air temperature and accumulated spring precipitation.
Station measurements of precipitation and air temperature are available through the National Climatic Data Center (NCDC) and NRCS.
PRISM, or the Parameter-elevation Regressions on Independent Slopes Model, comprises a gridded daily meteorological dataset. Values are derived as described in (Daly et al., 1994). PRISM data are useful for developing basin-average accumulated precipitation and air temperature, which may more accurately represent basin conditions than point measurements.
The Northeast Regional Climate Center (NRCC) also produces gridded precipitation and air temperature data.
The Pacific Northwest region's AgriMet program(AgriMet) produces station measurements and provides the weather data required to model evapotranspiration. Measurements include precipitation, air temperature, solar radiation, and wind speed.
Streamflow data are necessary both as the dependent variable and as a predictor dataset. According to (NRCS, 2011), streamflow can be useful as an indicator of fall baseflow entering the winter snow accumulation season and of streamflow response in early spring. Several state and federal agencies maintain streamflow databases.
U.S. Geological Survey (USGS) maintains the National Water Information System (NWIS), a network of around 1.9 million sites in the U.S.
Each of Reclamation's regions maintains its own streamflow database.
Great Plains Region-Reclamation typically forecasts inflows to its reservoirs. Great Plains region calculates daily reservoir inflow using a mass-balance approach, where inflow is equal to daily outflow minus daily change in reservoir storage. Therefore, calculated reservoir inflow is net gain, including evaporation and reservoir seepage.
Pacific Northwest Region-The Pacific Northwest region calculates daily reservoir inflow using a mass-balance approach, where inflow is equal to daily outflow minus daily change in reservoir storage. Therefore, calculated reservoir inflow is net gain, including evaporation and reservoir seepage.
The State of Colorado maintains a streamflow database here.
The Oregon Water Resources Department (OWRD) maintains a streamflow database.
The State of Wyoming maintains a database of surface water measurements. Data can be accessed through the State Engineer's Office.
Antecedent soil moisture can impact runoff (e.g., Penna et al., 2011). NRCS measures soil moisture through its Soil Climate Analysis Network (SCAN) at depths of 2, 4, 8, 20, and 40 inches. Because soil moisture can impact runoff, soil moisture measurements may prove useful for forecasting streamflow volumes.
A number of teleconnections impact weather in the Western U.S., and may prove useful for seasonal statistical forecasting. Indices of the El Niño Southern Oscillation have been shown to provide predictive skill in the Klamath basin in Oregon (Kennedy et al., 2009). Brief descriptions of teleconnections and indices pertinent to Western U.S. weather and climate appear below.
Indicators of the ENSO have shown value in forecasting in the Western United States (NRCS, 2011). ENSO has been shown to impact winter precipitation and air temperature, with general patterns shown in the figure below. Typically, La Niña years result in colder, wetter winters for the Northwest, and El Niño years result in wetter winters in Southwest and warmer winters in the Northwest. These impacts may result in snowpack accumulation trends for Western US basins.
ENSO impacts to U.S. winter jet stream and climate. From (NOAA, 2019).
The AO describes the state of atmospheric circulation over the arctic. According to the North Carolina Climate Office, positive AO is associated with warmer winters in the United States, whereas colder winters arrive with negative AO. Data are available through the Climate Prediction Center.
_ Images from National Geographic Magazine, March 2000; Sources: Doug Martinson, Wieslaw Maslowski, David Thompson, and John M. Wallace, via North Carolina Climate Office_
According to (Gottschalk, 2015) ...[The MJO] can have dramatic impacts in the mid-latitudes. Several times a year the MJO is a strong contributor to various extreme events in the United States, including Arctic air outbreaks during the winter months across the central and eastern portions of the United States. Similar to ENSO, the MJO occurs in the tropics. However, the MJO is transient whereas ENSO is stationary, and the MJO occurs over shorter periods than the seasonal ENSO.
PDSI is a general indicator of how wet or dry a region is over long-term periods. It is based on temperature data and a water balance model. According to Dai, Aiguo & National Center for Atmospheric Research Staff (Eds), 2017 "Monthly PDSI values do not capture droughts on time scales less than about 12 months."
Information regarding SPI can be found here.
Carroll, T., Cline, D., Olheiser, C., Rost, A., Nilsson, A., Fall, G., … Li, L. (2006). NOAA’S NATIONAL SNOW ANALYSES. 14. Daly, C., Neilson, R. P., & Phillips, D. L. (1994). A Statistical-Topographic Model for Mapping Climatological Precipitation over Mountainous Terrain. Journal of Applied Meteorology, 33(2), 140–158. https://doi.org/10.1175/1520-0450(1994)033<0140:ASTMFM>2.0.CO;2
Dai, Aiguo & National Center for Atmospheric Research Staff (Eds). Last modified 12 Jul 2017. "The Climate Data Guide: Palmer Drought Severity Index (PDSI)." Retrieved from https://climatedataguide.ucar.edu/climate-data/palmer-drought-severity-index-pdsi.
Kennedy, A. M., Garen, D. C., & Koch, R. W. (2009). The association between climate teleconnection indices and Upper Klamath seasonal streamflow: Trans-Niño Index. Hydrological Processes, 23(7), 973–984. https://doi.org/10.1002/hyp.7200
Lindsay, R. (2017, September 18). How El Niño and La Niña affect the winter jet stream and U.S. climate. Retrieved June 11, 2019, from https://www.climate.gov/news-features/featured-images/how-el-ni%C3%B1o-and-la-ni%C3%B1a-affect-winter-jet-stream-and-us-climate Natural Resources Conservation Service, U. S. D. of A. (2011). Part 622 Snow Survey and Water Supply Forecasting, Chapter 7: Water Supply Forecasting. In National Engineering Handbook (Amendment 41, p. 13). Retrieved from https://directives.sc.egov.usda.gov/OpenNonWebContent.aspx?content=32039.wba
Penna, D., Tromp-van Meerveld, H. J., Gobbi, A., Borga, M., & Dalla Fontana, G. (2011). The influence of soil moisture on threshold runoff generation processes in an alpine headwater catchment. Hydrology and Earth System Sciences, 15(3), 689–702. https://doi.org/10.5194/hess-15-689-2011