The National Drought Mitigation Center (NDMC) produces VegDRI in collaboration with:
Main researchers working on VegDRI are:
- Dr. Brian Wardlow, with the Center for Advanced Land Management Information Technologies (CALMIT)
- Dr. Tsegaye Tadesse at the NDMC
- Jesslyn Brown with the USGS
NDMC and CALMIT are both based in the School of Natural Resources at the University of Nebraska-Lincoln. VegDRI was developed with sponsorship from the US Geological Survey (USGS), the US Department of Agriculture's (USDA) Risk Management Agency (RMA) and from NASA. Continued operational production of VegDRI is supported by the USGS.
VegDRI maps are produced weekly and provide regional to sub-county scale information about drought's effects on vegetation.
In 2006, VegDRI covered seven states in the Northern Great Plains (CO, KS, MT, NE, ND, SD, and WY). It has undergone multiple expansions:
- 2007 - NM, OK, MO, and TX
- 2008 - WA, ID, OR, UT, CA, AZ and NV
- 2009 - Eastern U.S.
The VegDRI calculations integrate satellite-based observations of vegetation conditions, climate data, and other biophysical information such as land cover/land use type,
soil characteristics, and ecological setting. The VegDRI maps that are produced deliver continuous geographic coverage over large areas, and have inherently finer spatial detail
(1-km2 resolution) than other commonly available drought indicators such as the U.S. Drought Monitor.
1. What climate-related variables are incorporated into the VegDRI calculations?
- Palmer Drought Severity Index (PDSI)
- Standard Precipitation Index (SPI) (36-week)
Research is underway to test shorter-term drought indicators, such as the Palmer-Z
index and shorter SPIs (e.g., 4- and 8-week), as alternative climate-related variables.
2. What specific satellite-derived vegetation condition information is used in VegDRI?
Two variables related to general vegetation conditions – the Percent Average Seasonal Greenness (PASG) and Start of Season Anomaly (SOSA) – are calculated from satellite-based observations and incorporated into the VegDRI. Both variables are calculated from normalized difference vegetation index (NDVI) data acquired by NOAA’s Advanced Very High Resolution Radiometer (AVHRR), which acquires satellite imagery of the entire Earth on daily basis at a nominal spatial resolution of 1-kilometer2. NDVI is a dimensionless measure of the relative state and condition of vegetation from which metrics such as PASG and SOSA can be calculated to monitor and quantify seasonal vegetation dynamics. These metrics can be used to summarize the degree of variability in vegetation conditions from year to year. More information about NDVI and AVHRR can be found at: http://ivm.cr.usgs.gov/.
3. What is Percent Annual Seasonal Greenness (PASG)?
PASG compares the accumulated seasonal greenness (or NDVI) up to a specific point in the year to the historical average seasonal greenness (1989 to 2005) for that same date. A PASG value of 100% indicates that the current seasonal greenness is comparable to the average historical greenness for that location at the time. This would indicate that the vegetation conditions are near normal. PASG values less than 100% would indicate below average greenness (poorer than normal vegetation conditions) that may be linked to some form of stress (e.g., drought, flooding, fire, hail damage, or pest infestation). PASG values greater than 100% would indicate higher than average greenness, which would reflect better than normal vegetation conditions. PASG values are not calculated for a location until the start of season has occurred. Further information can be found at: https://vegdri.cr.usgs.gov/methods_vegdri.php
4. What is the Start of Season Anomaly (SOSA)?
The SOSA is the temporal difference in the start of season (SOS) for a given year compared to the historical average SOS for a location. SOS can be defined as the time at which vegetation initiates growth (photosynthetic activity) after winter as observed from satellite observations. (Note: There can be a delay between the SOS we observe on the ground versus the SOS that can be detected from satellite-based observations.) A negative SOSA indicates that the SOS for a specific year is later than the average date and a positive SOSA appears when green up occurs earlier than normal. The SOSA is used to express the inter-annual changes that can occur from year-to-year in the VegDRI model. Further information can be found at: http://phenology.cr.usgs.gov/methods_deriving.php.
5. Why not use satellite-derived NDVI information to monitor the level of drought stress on
vegetation instead of VegDRI?
NDVI has been widely used to monitor vegetation conditions, but this index reflects the collective influence of multiple environmental factors (climate, pest infestations or plant disease, land cover/land use change, fire, and extreme weather events such as hail and flooding) on the state and condition of vegetation. As a result, NDVI can be used to establish general vegetation conditions, but distinguishing the impact of drought on vegetation from the other factors can be done using only NDVI information. For VegDRI, the AVHRR NDVI data provides spatially-detailed (1-kilometer2) information on vegetation conditions (e.g., good or poor) across the US. Coarser-resolution climate data is used to identify which areas of stressed vegetation in the NDVI data are experiencing dryness, which distinguishes drought impacted areas from those influenced by other environmental stressors.
6. Why is biophysical information used in the calculation of VegDRI?
Information about soils, land cover, land use, and the ecological setting are incorporated into VegDRI because the climate-vegetation response can vary depending on these different environmental characteristics. The influence of drought stress on vegetation can vary depending by land cover type, soil characteristics, and land use practices. As a result, these variables were added to the VegDRI model to represent these environmental differences.
7. What specific biophysical parameters are analyzed in VegDRI?
Biophysical parameters in the current VegDRI model include:
- soil available water capacity (source: STATSGO database)
- land cover/land use type (source: USGS National Land Cover Dataset (NLCD) -2002).
- national irrigated lands map derived from 250-meter MODIS satellite data
- ecoregion type (source: EPA Omernik ecoregions Level 3)
- elevation (source: USGS National Elevation Dataset - 2002)
VegDRI Model and Maps
8. What does the "Out of Season" VegDRI category represent?
"Out of season" means that a signal indicative of photosynthetically-active vegetation has not yet been detected from satellite for that location during that specific calendar year. Out of Season represents the over-wintering period when the vegetation is dormant and crops are unplanted. As a result, no VegDRI value can be calculated until the growing season (as defined from satellite observations) begins. Out of Season length can vary depending on the vegetation type and geographic location. Also, conditions on the ground may show that the winter dormancy or the emergence of a crop has occurred but the VegDRI may still show that location to be "Out of Season". There is often an inherent delay (by a few days to week) between what is observed on the ground and what is detected from satellite. Identification of the start of the growing season from satellite requires certain amount of photosynthetically-active plant material be present before spectral signal indicative of actively growing vegetation can be detected in the NDVI data.
9. How many years of historical data are used to develop the VegDRI models?
The current VegDRI model is based on historical climate and satellite observations for a 20-year period spanning from 1989 to 2008. The length of the historical record is limited to this period because the satellite-record of 1-kilometer2 AVHRR satellite observations over the U.S. began in 1989. Ideally, a long historical record would be preferred but during this 20-year span, many locations have experienced representative events for most (or all) VegDRI categories. Prime examples are the extreme drought conditions throughout much of the Central U.S. in 2002 and the extremely moist spell in 1993 experienced by much of the U.S.
10. Are future updates or improvement of VegDRI planned?
Yes, we plan to investigate new model inputs and other changes to the VegDRI model over the upcoming year for the continental U.S. Changes will be made to the existing VegDRI model if specific inputs or changes correct or improve any major errors or weaknesses that are highlighted through evaluations of the VegDRI maps. The VegDRI maps produced in 2009 use the current model (VegDRI_4.0) and model improvements will be made before the 2010 growing season (VegDRI_5.0) based on user feedback and the results of our investigations. Subsequent updates may also be done in the future to improve the accuracy of VegDRI. As research results become available, they will be added to the VegDRI website for comment.
11. Will the temporal resolution of the VegDRI maps be improved?
The transition from bi-weekly VegDRI maps to weekly maps is planned in order to be consistent with other drought indicators (e.g., U.S. Drought Monitor) that are produced on that time step. No specific date has been set for the transition to a weekly VegDRI product, but this activity is planned for the near future.
12. Will a historical time series of VegDRI maps be available dating back to 1989?
Yes, the VegDRI maps will be calculated from 1989 to present once a stable, accurate VegDRI model is developed.
13. How is the accuracy of VegDRI assessed?
No single measure can be used to assess the accuracy (both quantitatively and qualitatively) of the VegDRI because of the varying definitions of drought. Assessments of other commonly used drought indicators such as the U.S. Drought Monitor (USDM) ( http://www.droughtmonitor.unl.edu/) also face the similar challenges. As a result, a variety of information sources will collectively be reviewed to evaluate VegDRI’s performance and accuracy. Periodic feedback from experts (e.g., state climatologists, USDM authors, and agricultural experts), agricultural producers, and others in the general public will be used to characterize the general strengths and weaknesses of VegDRI and highlight specific locations or trends that might be in error. We will also quantitatively compare the VegDRI to USDA crop yield data as well as, biophysical and soil moisture measurements being collected by various agencies and organizations (where available). The spatial and temporal patterns in the VegDRI maps will also be qualitatively compared to the drought patterns in the USDM maps and to the spatial distribution, type, and frequency of drought impacts being reported in the Drought Impact Reporter (DIR) ( http://droughtreporter.unl.edu/).
14. Will other VegDRI informational products be produced in addition to the maps?
Yes, the U.S. Geological Survey (USGS) hosts a Drought Monitoring Viewer ( http://gisdata.usgs.gov/website/Drought_Monitoring/viewer.php) that displays the VegDRI information in a dynamic viewer that allows users to zoom to a local view of the VegDRI, overlay and compare the index with multiple data layers (e.g., political boundaries, roads, rivers, and elevation), and print maps over any geographic extent. New experimental products will also be made available during 2010 on VegDRI website ( http://vegdri.unl.edu/) hosted by the National Drought Mitigation Center (NDMC), which include:
- animations of VegDRI maps across the growing season to visualize the changing drought patterns across a specific region, or state.
- VegDRI change maps (change in VegDRI values between the current date and a previous date or historical average),
- tables that summarize the total area covered by each VegDRI drought categories for a defined geographic area (i.e., state or county),
- trend lines that plot the VegDRI trend for a given year(s) over a specific geographic area (e.g., state or county) and a specific land cover type or types
Additional experimental VegDRI products may be produced based on user feedback.
15. What research activities are planned for VegDRI?
Several new or improved inputs for VegDRI will be investigated, which include:
- 1-km Land Surface Temperature (LST) data from the Moderate Resolution Imaging Spectroradiometer (MODIS). Day and night LST observations and derivative products will be investigated as inputs into the VegDRI model and also, for their ability to characterize "flash" drought events.
- Normalized Difference Water Index (NDWI) calculated from MODIS 500-meter near- and mid-infrared spectral reflectance data. The NDWI responds to both changes in the water content and intra-cellular air spaces of a plant’s leaves. As a result, it may respond more rapidly than NDVI to drought-induced stress events in vegetation such as wilting and desiccation. The initial results from this research are presented in a paper by Gu et al. (2007), which can be downloaded here.
- Rapid crop classification will be investigated using time-series MODIS 250-meter vegetation index (VI). Cropland areas are temporally static in the current land cover data set being used in the VegDRI model and inter-annual crop rotations are not reflected in the model. This research will attempt to develop methods to rapidly map crop types with different seasonal behaviors (e.g., winter wheat – spring growth cycle, summer crops – summer cycle, and alfalfa – cycle spanning spring, summer, and fall) on an annual basis. Such information would be used to better represent these land cover change dynamics over cropped areas in the VegDRI models.
- Crosswalk of VI data sets between the AVHRR and MODIS sensors. The crosswalk will attempt to make the AVHRR and MODIS NDVI values compatible and extend the current MODIS NDVI time series history (2000 – present) back to the beginning of the 1-km AVHRR NDVI historical record in 1989. The MODIS NDVI data have considerably higher quality than AVHRR NDVI data (particularly spectral hand placement, radiometric resolution, atmospheric correction, and geolocation accuracy), which will improve our large-area vegetation monitoring capabilities. The crosswalk will enable MODIS NDVI-based models to be generated from 18 years of historical information rather than 8 years, which will result in more accurate model results. Future research will also be undertaken to crosswalk the AVHRR-MODIS NDVI series to NOAA’s operational VIIRS (Visible/Infrared Imager/Radiometer Suite) NDVI data stream when the instrument is launched as follow-on to MODIS.
- 250-meter VegDRI The development of a 250-meter VegDRI will be explored using 250-meter MODIS Vegetation Index (VI) data. A higher spatial resolution VegDRI would be more applicable for local-scale monitoring and decision making than its 1-kilometer2 counterpart. The initial research will rely on shorter, 8-year MODIS historical record, which will be extended once the AVHRR-MODIS crosswalk activity is completed.