Oregon IPM Center at Oregon State University IPM PIPE The PRISM Group at OSU Western
Region Pest Management Center National Institute of
Food and Agriculture

IPM Weather Data, Degree-Days, and Plant Disease Risk Models
for agricultural and pest management decision making in the US

at Oregon State University Integrated Plant Protection Center's


1. Introduction
1b. Whats new page
2. Description and format of the data
3. Using and downloading data
4. Suggested uses for the data
5. Degree-day calculation instructions
6. Virtual hourly data - estimated from nearby stations and elevation
7. Missing data - Max Min data from virtual hourly data
8. Missing data - PRISM climate map interpolation
9. NOAA NMME extended weather forecasts
10. NOAA CFSv2 extended weather forecasts
11. Use prior years data as a forecast.
12. Missing data - 10 and 30 Year Histor. Avg Weather Data Interpolation
13. Missing data - time average interpolation
14. Weather forecasts (Fox Weather LLC and IPPC)
15. Weather forecasts (NWS)
16. Disclaimer

In cooperation with the Oregon State University, Oregon Climate Service, the OSU PRISM Data Group, APHIS PPQ, the National Plant Diagnostic Network, the Regional IPM Centers, and many others, the Integrated Plant Protection Center at OSU has has been working to make real-time daily weather data, degree-days, phenology models, and plant disease risk models available to serve agricultural pest management, crop management, and natural resource management needs for the US and southern Canada using a variety of weather networks that can be shared publicly. These formatted weather data, degree-days, degree-day products (calculators, models, maps, and map calculator), and hourly weather driven models are kept current and relevant for pest management decision making purposes for Extension personnel, growers, field persons, consultants, researchers, and students.

See Also:
List of Degree-Day Models
List of Hourly Weather Driven Models and Their Documentation

NEW: Email "Push" Notification Account Sign-up

Our sub-hourly, hourly and daily weather data come from several sources: the US Bureau of Reclamation AGRIMET and HYDROMET weather networks, the National Weather Service ASOS/METAR and COOP networks, the BLM/USFS RAWS network, the NRCS SNOTEL network, the CWOP/APRSWXNET network, and numerous others including these grower/agricultural networks: WSU AgWeatherNET (which does not allow 3rd party sharing of raw data), CPS Adcon Networks, IFPNet Automata, California CIMIS, California PestCast, and others. Most data are ingested from MesoWest in Utah.

All data are electronic and run automatically, usually with 15 minute updates. Currently we have 25,000+ stations total among all US states and surrounding areas.

If you have a weather station that you would like to share publicly, consider joining CWOP/APRSWXNET and go to this website to do so: www.wxqa.com/SIGN-UP.html

You can view the full list of all locations at uspest.org/data/sta_list.txt or, alternatively see, interact, and calc DDs using a map of stations by clicking this image:

Base 32 DDs linked to calculator (select Region and click map)

Real time (current) data feeds for many of these stations are linked from either the state, local, or region tables (Oregon, Medford OR, Hood River OR, WA, ID, MT, WY, N. Central, Great Lakes Central, NE, SW, S. Central, SE, Alaska, and SW Canada) or from this NWS Mesowest page.

We use computer programs that take these reports, sort and reformat them, calculate daily degree-days (50 F lower threshold), and append them to the appropriate files stored at several computers related to this project. All sources of data are sometimes lost due to station, computer and satellite downtime. Because 100% accuracy is less important than is consistency of the data, we use several data interpolation/estimation schemes (described below). The precipitation data are less accurate and we do not currently support a means to estimate missing data, other than to assume "zero" values.

The the format of the dayfile (max-min) data are:

          Temp.   Temp.
Month Day Max     Min    Rainfall  Degree-days
          deg. F  deg. F inches    Tlow=50 F
  1     1  55      37     0.050    2.03
  1     2  45      34     0.000    0.00
  1     3  58      32     0.120    4.39
  1     4  52      27     0.000    0.03
  1     5  49      30     0.000    0.00
 etc... more detailed format info 

We are hosting several forms of data: (all described below)
1. Current data from this year (updated daily)
2. Historical data from recent years (data archives)
3. Historical average temperature data (1981-2010, and recent 10-yr avg)
4. 7-day Forecast data from Fox Weather and NOAA NWS
5. Extended Forecast data (up to 7 months derived from NOAA NMME and NOAA CFSv
6. Hourly and sub-hourly data for plant disease risk models

First, note that this site is a secondary source of data: please go to the National Climate Data Center (NCDC) or this NCDC data access page, or the Agrimet data access page, or any of the other networks linked above as your primary sources of recent and archival data.

Our website provides an integrated alternative for those who mainly need to calculate degree-days, and run pest, crop, and plant disease models, but, you may want to use data from the primary sources above if you want the best accuracy. To download or view our data, from the station list map (see link above), or state/region tables (links above and on homepage), find the weather station nearest your area of interest. If using the state/region tables, to download data, click on either the range of years under "30 year average" or a year (such as 2018) for that location. For example click 2018 for Corvallis from the Oregon state table. After you have "browsed" to the data, your web browser should have an option to allow you to save the file to your local disk drive, in whatever directory you select. If using Netscape, the command is "File", "Save as" in the main menu. Then you specify the local drive and directory where you wish to save it. Once you save the data, you may load it into a graphics plotting program (and most spreadsheet programs) to make your own charts and do extended analyses and comparisons. We have tried to keep the formatting of the results minimal to facilitate the ASCII (text) importing feature that virtually all charting programs include. Refer to your program documentation for more details.

There are data from thousands of stations on-line at this time. More sites may be made available in the future. Contact Len Coop at coopl@science.oregonstate.edu to request other sites.

First, this website is intended to supply near real-time access to formatted temperature and other weather information to inform pest management decision making in the US. Various pests in different crops have been researched, and their development can often be predicted using degree- days [visit the degree-day faq page for definitions and more info]. At Oregon State University, we are compiling known information about certain pest species concerning development, threshold temperatures, and expected degree-days for the appearance of different life stages, and other events such as risk to disease infection. We present this information in the form of tables and charts. These tables may be reviewed from this list: degree-day/phenology models. This system will flag special events in the life-cycles of these organisms at specified degree-days after a starting point (usually called a biofix, at least for insect models). A file-upload feature of ddmodel.pl allows anyone to create daily max-min temperature data files from their own data, and submit them. This calculator also has an option for making graphs of the current vs. historical average degree-days, with events plotted.

You may also access the calculator with fewer options from any of the weather station tables such as Oregon current data. These allow you to enter your own upper and lower thresholds and calculation method, and link to forecast or historical data for predicting future degree-days. You also may download and save these data to make your own charts, typically using a spreadsheet program.

If you wish to compare the progress of degree-day accumulations with either historical normals or between regions, you may want to visit the US degree-day maps webpage. There you will find degree day maps with cumulative degree-days from Jan 1 to the current date, historical normal degree-days over a comparable interval, and deviations between the two. These maps represent the use of GIS (geographic information systems) to interpolate missing data between sites. These map are available for all coterminous US regions, and have both Google Maps and GRASSlinks GIS web interfaces to query degree-day values and to access the degree-day calculator.

We have several related versions of degree-day calculation tools. The next several sections generally describe how to use them. If you see only form fields for thresholds and calculation method, you may enter these manually. If you are using the full-featured DD model, you may find the exact model you need in the pull-down list; it can be a time-saver and a convenient way to automate much of the chore of using degree-days. The model database stores thresholds, calculation methods, starting dates, brief biofix instructions, and selected look-up events for each species model. But there are hundreds or more uses for degree-day calculations, so we allow you to enter your own parameters to use the tool in calculator mode. You may want to review UC Davis's Database of IPM models for a large collection of research summaries of phenology models for insects, mites, diseases, plants, and beneficials. From these and other sources, you will determine the parameters to enter.

CALCULATION METHODS - See also actual calculations and frequently asked questions on degree-days
A degree-day is a heat unit calculated on a daily basis, usually using the daily max and min temperature. Not all degree-days are equivalent or are directly comparible. Various formulas are used that produce different results. We can also vary the lower threshold, use or not use an upper threshold, and vary that as well. Although the online tools support several types of degree-days, there are several that are not yet supported. For example, a "vertical cutoff method" is sometimes used for the upper threshold. We have not implemented the vertical (or intermediate) upper threshold cutoff methods, only horizontal. If you are not certain what type to use, we suggest trying the most common Simple Average or Single Sine methods (with no upper threshold), or better yet go back and find your reference documentation to determine which method to use. We use Single Sine as the default method following the UC Davis recommendations.

In short, we have:
Simple Average - this is the daily max plus min divided by 2 minus the lower threshold (TLOW). This is commonly known as a "growing degree-day"; a term we do not use at this website due to the slightly different "corn growing degree-day".

Growing DDs (GDDs) - this is the formula widely used for sweet corn. We calculate GDDs for corn by using the lower threshold (TLOW) (50 deg. F) if the min is less than the TLOW, and by using THI (86 deg. F) if the max is greater than the THI, computing the average and then subtracting the TLOW. NOTE that for cereals, grapes, and many crops, the term "growing degree day" is used for what otherwise are called simple average degree days: (max + min)/2 - TLOW
Also, note that "growing degree-day" is the common term for "simple average degree-days" in many regions of the USA. We do not endorse this usage. Insect development (phenology) is largely determined by heat units (degree-days). Insect GROWTH is a different concept - growth depends upon food availability and other factors to a much greater extent than development. Since corn GDDs were in use in earlier times than insect DDs, we retain the historical definitions.

Single Triangle, Double Triangle, Single Sine, and Double Sine - these formulas are more complex and somewhat more precise, and are well described at the UC Davis IPM website.

Heating and Cooling - these are used in the energy and power industries to compute estimates of fuel consumption and are provided here as a convenience. These are generally not used for pest management. Note that an advanced use of heating degree-days is to estimate chilling or frost units, which can be used to estimate insect population mortality rates during winter periods.

Potato P-days - these were developed by Sands, P.J. 1988. A pocket-calculator model predicting graded yields and development in potatoes. Potato Research. 31:521-534. They are used by the potato early blight model.

The lower temperature threshold (TLOW) is the theoretical temperature at which development begins to respond to temperature. This is usually determined either in the laboratory, from constant temperature rearing studies, or empirically from a series of data from the field. Common thresholds for insects range from 41 to 50 F, and for plants usually between 32 and 45 F. Upper thresholds are not used for most insect and crop models. If your documentation describes the use of an upper threshold, be sure that it is of the "horizontal cutoff" or "intermediate cutoff" type. Do not use this calculator if you need the "vertical cutoff" type of upper threshold; we currently do not support it.

If you have selected a model (not calculator mode), then the thresholds and calculation method will be displayed for that model. Do not change these settings. If you wish to change them, then they will be ignored unless you are in calculator mode.

The starting date for DD calculations is critical, the ending date is usually not (it's OK to run past the normal season). Enter the starting date according to instructions if known. In many cases, the starting date is fixed to a calendar date. For others, you may need to trap or sample the organism to know what time the biofix/starting date begins. Refer to any available model documentation for details.

The degree-day calculator and modeling tool has the ability to store settings in the web address window of your browser, which can then be saved as Bookmarks (Netscape/Mozilla/Firefox/most other web browsers) or as Favorites (Internet Explorer) for future use. This can be a major time saver to those who hate filling out the same form over and over. As a convenience and time saver, you may wish to make one or more bookmarks (or weblinks on your own webpages) to store your individual settings. This feature has been improved and simplified so that nearly all settings can be stored in this way.

To do this, simply click the checkbox underneath the model pulldown list after you make all of your settings such as the thresholds, dates, location, and options. (the place to click is shown here underneath the model pulldown list):
click the box shown here after setting all your form elements
Once you have filled out the form with all your settings, then click the checkbox. After clicking this checkbox, your URL/web address should now be displaying all your settings, for example like this:
example URL after clicking the checkbox showing your settings
Then, either immediately, or after running the calculator (click the calc button), you may save the settings by making a bookmark (in Mozilla/Firefox click the "Bookmarks", then "Bookmark this page", in Internet Explorer click "Favorites", then "Add to Favorites").

If you routinely run several different models or the same models but with different settings, you will want to repeat this process and to provide special names (or titles) for each group of settings. For example in Internet Explorer, each time you click "Add to Favorites" a box will open that includes a "Name:" field. Edit that field to describe your current settings to distinguish it from other settings that you want to save. With other web browsers, a similar naming of bookmarks should be possible.

You may also make weblinks from your own web pages using these settings (useful for Extension agents who make custom webpages for their local users). Currently only the "file upload" setting cannot be saved using bookmarks.

This feature should ease the tedium of re-entering your calculator/model settings and encourage you to run the program more frequently.

Data outages are common, and various estimation methods and forecasts are available depending on the application. We briefly introduce them and the tags used to indicate them in our output programs in this section:

6. VIRTUAL HOURLY DATA - Estimated from nearby stations and elevation
What do the "*vdata*" tags mean?

Beginning in March 2011, web users may log in and register their own "v2 virtual weather stations" to run models for locations that may not be near actual weather stations carried on our system. We are adding this tag to degree-day models to show that these virtual data are running the model. This is a new feature being tested starting in 2011, and it must be used with caution. In particular, if any nearby weather stations are poorly sited or have mis-calibrated temperature sensors, then these data may be poorly estimated. Our data QA/QC programs attempt to filter out such problem data, but slightly or moderately incorrect data can be easily overlooked and create a local bias for your virtual station. Therefore, please notify us if you believe there are any problems with the virtual data. In any case, always verify that the data look reasonable, are similar to actual *quality* nearby stations, and perform according to your needs and expectations.

The algorithm used to estimate the data can be described as distance-weighted, elevation-regression of nearby station data. This regression analysis uses elevation to create a "local lapse rate" equation for every temperature, dewpoint, and precipitation estimate for hourly virtual data. Virtual daily max-min data used to run degree-day models is created by extracting it from the hourly data.


7. MISSING DATA - Max-Min data from virtual hourly data
What do the "Vx" and "Vn" tags mean?

When the PRISM map interpolation method is not available (below), we attempt to use our virtual hourly weather data to fill in missing values from the observing weather station. This is a method we developed for our hourly weather data QA/QC that we are beginning to use for missing max/min data as well. The method, briefly, uses a distance- weighted elevation-regression approach to using nearby station observations to fill in missing data for the selected station. We name these stations by adding "v2" in front of regular station names, e.g. "v2CRVO". For more information see "VIRTUAL HOURLY DATA" above.


8. MISSING DATA - PRISM Climate Map Interpolation
What do the "Px" and "Pn" tags mean?

A PRISM "CAI" (see below) climate map-based missing data interpolation has been in use here since 2003. This is a rather calculation-intensive way to estimate missing temperature values. It is very similar to the method used to generate degree-day maps. Essentially we compute daily max and min temperature maps (using our real-time data and PRISM climate maps to improve surface interpolations), and then use these maps to estimate values for any missing days for all temperature data files.

This "climate map" interpolation method has the advantage that it considers elevation and local terrain in estimating values from nearby stations, and that consecutive dates of missing data can be better estimated than when using the "time average" method below. Plus, if one or more "good" weather stations are fairly close, then estimates should be fairly accurate and unbiased.

If you see a small number of these tags in the model/calculator output, then the errors in degree day predictions attributable to missing data should be small and of relatively little concern. However, if many of these tags are displayed, then the models may be in substantial error, and you should consider selecting a different location or using your own data via the file upload feature of the model. Please be careful in trusting the model under all conditions, especially when the missing data count is high.


9. NOAA NMME 7-month Extended Seasonal Forecast
What does the "Nm" tag mean?

Starting Feb 2016 we have a new 7-month seasonal "climate" forecast that uses NOAA's North American Multi-Model Ensemble (NMME) forecast system, which allows us to use a science-based model forecast to extend our model predictions for up to 7-months into the future, based on an ensemble of 7 different long-range numerical forecast models.

We use the gridded NMME 1-7 month forecasts of monthly climate anomalies, PRISM climate data, and temporal interpolation to daily max, min, and precipitation values for this new feature. This should be a useful alternative to 10 and 30 year historical average weather data.

For example here are links to the 1-7 month N. American NMME forecasts for temperature and precipitation anomaly maps.

This NOAA ensemble forecast system is documented in: Kirtman et al. 2014. The North American Multimodel Ensemble. Phase-1 Seasonal-to-Interannual Prediction; Phase-2 toward Developing Intraseasonal Prediction. Bull. Amer. Meteorol. Soc. 585-601.


10. NOAA CFSv2 3.5-month Extended Seasonal Forecast - NEW extended forecast option
Starting Oct 2016 we have a second new seasonal forecast - currently out 3.5 months, similar to NMME above except for these differences:

  1. This is a single model (not an ensemble) known as the NCEP Couple Forecast System model version 2 from NOAA.
  2. The interpolation approach taken was to attempt to reconstruct reasonable "rainfall patterns" or episodes to emulate actual weather rather than to smoothly interpolate temperatures and rainfall amounts as was done with our NMME seasonal forecast.
  3. The work is done by our collaborator Fox Weather LLC rather than by OSU IPPC.
  4. The process only goes out for 3.5 months rather than 7 months.
  5. The workflow may not stable.
For either extended forecast (any forecast for that matter), use with caution.


11. Use prior years data as a forecast.
To round out our options for model forecasts, we allow access of data from either of the 2 previous years (if available). This option may make sense if the current year thus far is similar to either prior year. One way to check this is to look at the current "Date Comparison" report. If the selected location is currently only a few days different from a prior year, it may support the use of that year as a forecast. Our Highcharts graphs include all 6 forecast types in the projected degree-days.


12. MISSING DATA - 10 and 30 Year Historical Average Weather Data Interpolation
What do the "Hx", "Hn", and
PRISM-CAI tags mean?

New Aug. 2015 - 10-year avg data used for extended forecasts in many (not all) degree-day calculation products.
Hx and Hn are the recent 10-year historical average data adjusted for the local site (using the station code location), used both to fill in missing values (if PxPn and VxVn data are not available, see above), and for extended forecasts. Starting in 2015 we developed a "rolling" 10-year average weather data type. This simply uses daily PRISM temperature grids averaged over the past 10 years for Tmax, Tmin, and Precip. each day. We then sample the average PRISM grids at each weather station location. These data are updated monthly. The older (not currently in use) 30-year data are computed from nearby 30-year normal stations+PRISM data using GIS (PRISM-CAI means "climatologically aided interpolation using PRISM climate maps") for each observing station location.


13. MISSING DATA - Time Average Weather Data Interpolation
What do the "Mx" and "Mn" tags mean?

With this method (should all of the above not succeed), if one days' data are missing, we average data from the day before and the day after. The "Time Average" method has the advantage of being 1) simple to implement, and 2) relatively accurate at least if only one day is estimated with this method. This is now our "last resort" method for 48-state CONUS data. For non-CONUS, this may be the ONLY method available; so be wary if more than several days are missing (indicated by the Mx and Mn tags).

GENERAL NOTES about missing weather data:

The occurence of missing data (as indicated by these various tags) varies considerably depending upon the network and systems involved in delivering the data. In general, the Agrimet/Hydromet and METAR (National Weather Service via Mesowest Utah & Missoula Montana) have been among the more reliable networks for delivering consistent, reliable temperature max and min values.

If you see a small number of these tags in the model/calculator output, then the errors in degree day predictions attributable to missing data will be very small and of little concern. However, if many of these tags are displayed, then the models may be in substantial error, and you should consider selecting a different location or using your own data via the file upload feature of the model. Please be careful in trusting the model under all conditions, especially when the missing data count is high.


What do the "Fx" and "Fn" tags mean?

Starting in 2012 (and ending in 2021), we contracted with Fox Weather, LLC to build a forecast server customized for IPPC, using tools funded by USDA NIFA grants headed by IPPC, thus the forecast is owned by both Fox Weather and IPPC. The actual forecast is based on the NOAA NWS GFS (Global Forecast System) model, downscaled to 12 km using the WRF-ARF weather model, and further downscaled to 1.5 km by Fox Weather MtnRT(c). This results in a custom forecast that can be configured for multiple agricultural needs, covering all western CONUS states west of the 105 degrees West longitude. This provides an alternative to using the NOAA NDFD forecast in Western states.


What does the "Nd" tag mean?

Starting in 2013, we began using the 2.5km NOAA NDFD gridded and real-time forecasts. If gridded, we sample daily CONUS grids for all stations in our station list, and use those in many of our products. If that fails, we may revert to either NDFD real-time (gathered on-the-fly from NWS) or from our legacy NWS zone forecast system (described below).

Starting in 2004 (ending around 2016), we also used the National Weather Service IWIN (Interactive Weather Information Network).


No longer available.


All data and products are provided "as is" and users assume all risk in their use. No claims are made as to the correctness or appropriateness of this information for your particular needs. No specific pest control products are intended for endorsement or use. These responsibilities and all associated liabilities rest solely with the people who interpret and implement information from this and other sources. Use all predictive information with caution - errors occur, and predictive models do not replace the need for proper monitoring in the field. If you observe conditions that differ substantially from model predictions, please contact us to determine if the model inputs were incorrect, if the model functioning or weather data are in error, or if the model is inappropriate for your conditions.

[Home] [Intro] [DD Models & Calc] [Links]

This project funded in part by grants from the USDA NIFA, USDA APHIS PPQ, USDA SARE, USDA-Western Regional IPM Center Signature program, and from Oregon Statewide IPM funds.

On-line since April 5, 1996
Last updated Jan. 6, 2022
Contact Len Coop at coopl@oregonstate.edu or 541-737-5523 if you have any questions about this information.