1 / 39

Greg Easson, H. G. Momm The University of Mississippi Ronald Bingner

Evaluation for the Integration of a Virtual Evapotranspiration Sensor Based on VIIRS and Passive Microwave Sensors into the Annualized Agricultural Non-Point Source (AnnAGNPS) Pollution Model. Greg Easson, H. G. Momm The University of Mississippi Ronald Bingner

kenton
Download Presentation

Greg Easson, H. G. Momm The University of Mississippi Ronald Bingner

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Evaluation for the Integration of a Virtual Evapotranspiration Sensor Based on VIIRS and Passive Microwave Sensors into the Annualized Agricultural Non-Point Source (AnnAGNPS) Pollution Model Greg Easson, H. G. Momm The University of Mississippi Ronald Bingner USDA – ARS – National Sedimentation Laboratory

  2. Project Objectives • To investigate the feasibility of using existing NASA results as source of ET estimates for AnnAGNPS pollution model • To evaluate the continuity of the NASA-based remotely sensed ET estimates by the future missions

  3. Project Rationale • Evapotranspiration (ET) plays an important role for modeling surface-lower atmospheric flux processes • ET estimates in a continuous and spatially distributed fashion represents a challenge for scientists • Remote sensing-based techniques are sought as an possible alternative

  4. Background: AnnAGNPS • The Annualized Agricultural Non-Point Source • Pollution model is a continuous watershed-scale computer simulation tool used to generate loading estimates for some constituents of agricultural non-point source pollution

  5. Background: AnnAGNPS (continued) • Developed by USDA-NRCS • Event driven model • Simulates • Surface flow • Sediment • Nutrients • Pesticides • Used to evaluate Best Management Practices

  6. Background: AnnAGNPS (continued) • Watershed is divided into cells • Each of these cells requires 22 parameters • Climate data is derived from field weather stations located within or nearby the watershed • Thiessen polygon method

  7. Background: AnnAGNPS (continued) • Problem when field weather stations are sparse or even non-existing

  8. Project Objectives • To investigate the feasibility of using existing NASA results as source of ET estimates for AnnAGNPS pollution model • To evaluate the continuity of the NASA-based remotely sensed ET estimates by the future missions

  9. Evaluation of the Integration of NASA Results into AnnAGNPS • Modifications to AnnAGNPS • Concept of “Virtual” field weather stations

  10. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Modifications to AnnAGNPS

  11. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Study Site Long history of hydrologic work Extensive infrastructure USDA-ARS NSL past and ongoing projects

  12. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • MOD16 daily images for 2004 • Provided by scientists at The University of Montana (Nishida et al., 2003, Cleugh et al., 2007, and Mu et al., 2007). • Ground sampling distance (GSD) of approximately 5,000 meters

  13. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Two AnnAGNPS simulations • ET computed using the Penman equation • ET provided proxy-MOD16

  14. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Results: • Average watershed ET

  15. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Results: • Daily runoff

  16. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Results: • Spatial distribution of the 2004 annual percent difference between ET from AnnAGNPS and from MODIS

  17. Evaluation of the Integration of NASA Results into AnnAGNPS (continued) • Results: • Spatial distribution of the 2004 annual percent difference between runoff from AnnAGNPS and from MODIS

  18. Project Objectives • To investigate the feasibility of using existing NASA results as source of ET estimates for AnnAGNPS pollution model • To evaluate the continuity of the NASA-based remotely sensed ET estimates by the future missions

  19. Comparison of Existing and Future NASA Results • Due to the lack of published methodology describing the generation of ET estimates from VIIRS data, a different approach was considered • Using the relationship between ET, VI, and LST, daily ET maps were generated from models created using multivariate linear regression techniques

  20. Comparison of Existing and Future NASA Results (continued) • Lambin and Ehrlich’s feature space

  21. Comparison of Existing and Future NASA Results (continued) • Daily images from April 01, 2004 to July 31, 2004 • Re-sampled to 5,000 GSD 250 meter MODIS NDVI pixels 1,000 meter MODIS LST pixels 400 meter proxy-VIIRS NDVI pixels 750 meter proxy-VIIRS LST pixels

  22. Comparison of Existing and Future NASA Results (continued) • “Virtual” stations “Virtual” Field

  23. Comparison of Existing and Future NASA Results (continued) • Simplified representation DOY 1 DOY 2 DOY 3 DOY 4 5 5 5 5 1 2 1 2 1 2 1 2 3 4 3 4 3 4 3 4 1 2 3 4 5 4 5 2 3 1 DOY 3 4 5 1 2 4 5 1 2 3 Stations

  24. Comparison of Existing and Future NASA Results (continued) • Simplified representation

  25. Comparison of Existing and Future NASA Results (continued) • Model development • Stations 127 to 136 (physical stations) • Stepwise backward elimination (P-value associated with Pearson’s Chi-Squared). • One model per day for each of the sensors considered

  26. Comparison of Existing and Future NASA Results (continued) • Adjusted R2 > 0.25

  27. Comparison of Existing and Future NASA Results (continued) • Results • Variability of models performance • Adjusted R2 • Predictors

  28. Comparison of Existing and Future NASA Results (continued)

  29. Comparison of Existing and Future NASA Results (continued) • Simplified representation

  30. Comparison of Existing and Future NASA Results (continued)

  31. Conclusions • Linking MODIS ET with AnnAGNPS was successfully performed. • The use of MODIS ET can reduce the need to collect/generate dew point, wind speed, and cloud coverage.

  32. Conclusions (continued) • Reducing uncertainty in input parameters will reduce the uncertainty in the model results. • In addition, these values usually have temporal and spatial variability that are not easily taken into consideration when computing ET values.

  33. Conclusions (continued) • MODIS-ET produced 35% less ET then AnnAGNPS-ET and resulted in a 10% increase in runoff. • Large watershed system, climate parameters can be highly variable.

  34. Conclusions (continued) • MODIS-ET provided a more comprehensive spatial variability capability than is not often available from measured climate stations. • Additional remotely sensed data: precipitation and temperature.

  35. Conclusions (continued) • The second objective of this research project was to investigate the continuity of future NASA missions in providing ET estimates to AnnAGNPS simulation model. • Daily NDVI and LST maps from MODIS and proxy-VIIRS data were used to create two sets of daily ET maps.

  36. Conclusions (continued) • Direct comparison between these two sets of daily ET maps indicates that the next generation of moderate resolution sensor will continue to be a potential source of ET estimates to simulation models such as AnnAGNPS. • The VIIRS’s physical design features, such as improved signal to noise ratio and the attenuation of the “bowtie-shaped” footprint at large scan angles were not considered.

  37. Conclusions (continued) • The spatial variability demonstrated by the VIIRS-based LST map can be in part attributed to the downscaling technique used in the simulation process. • Further investigation should be conducted to estimate ET for different land use/land cover classes.

  38. Conclusions (continued) • There are situations were the ET maps generated from VIIRS and from MODIS agreed. • This demonstrates the potential of VIIRS to be used as the continuity mission, in providing ET estimates for AnnAGNPS pollution model.

  39. Acknowledgements • Institute for Technology Development • National Sedimentation Laboratory • The University of Montana • NASA and the University of Southern Mississippi

More Related