INTRODUCTION
We are developing a series of degree-day accumulation maps that can be used for a number of purposes such as:

• Comparing different growing regions for relative heat unit accumulations
• Showing average heat units for prediction of upcoming events
• Comparing this year vs historical average heat unit accumulations

These maps are primarily based on the relationship between temperature and elevation. As you go higher up, temperatures decrease. This relationship is fairly consistent for a given region, but will differ significantly between regions. The PRISM model is an expert system approach to determining local relationships between climate and elevation. It was developed at the Oregon Climate Center at Oregon State University and has its own web site.

The PRISM model uses a state-wide spatial resolution of ca. 30s (800m), meaning that the length of a cell (a "pixel" of data) border is ca. 800meters. After maps are computed, we use a simple 2-pass gaussian filter to "smooth" the data. This does not improve precision of the data, but it does improve its appearance somewhat.

On computing historical NORMAL maps: We start with PRISM-built maps for CONUS monthly maximum and minimum temperatures that are on-line and available from the PRISM FTP site. At the same time, we use historical average weather data from hundreds to thousands of sites in CONUS, and calculate cumulative degree-days. Using a GIS (GRASS), we next compute degree-days from the PRISM data and found the difference between the two data sources. We then perform an inverse-distance weighted interpolation of the differences, and added these values to the PRISM-derived degree-days. TIGER (US Census bureau) highways and county boundaries were added for reference.

On computing daily updated maps: For the present year maps, we use all available sites within major networks (ASOS/METAR, RAWS, AGRIMET, others). Each day after the weather data has been updated, we re-calculate cumulative degree-days. We use the Single Sine method for sites and Single Triangle method for uncorrected PRISM maps. If a site has bad or missing data, that site will be ignored until the data are corrected. As with the NORMAL maps, we find the difference between the PRISM degree-days and the actual degree-days for each location. We then interpolate the differences and add these values to the PRISM maps.

On computing daily deviations between historical NORMAL and actual data: Once the two map types have been re-calculated, we simply subtract the NORMAL maps from the actual maps, to obtain the DEVIATIONS from NORMAL. If it was 200 DDs cooler this year than NORMAL for a given location, it will be indicated using a dark grey scale (black in this case). For areas where it is warmer than NORMAL, the same color legend that is used for actual degree-days is used.

Preliminary Metadata Summary:
• Data Origin: Interpolated CONUS site calculated degree-days (32, 41 and 50 F base threshold) from multiple sources [need to update] . See also PRISM data description.
  
• Spatial Resolution: 30 arc seconds (ca 800m).
  
• Temporal Resolution: PRISM data- monthly average daily maximum and minimum temperatures used. Site data- daily maximum and minimum terperatures.
  

On-line since April 5, 1996
This page last updated Jun 26, 2017
Contact Len Coop at coopl@bcc.orst.edu if you have any questions about this information.