Remote Sensing in Drought Monitoring: Myths and Realities

1
Remote Sensing in Drought Monitoring Myths and
Realities

• Jesslyn F. Brown
• SAIC, contractor for USGS at the Center for Earth
  Resources Observation and Science (EROS)
• 5th U.S. Drought Monitor Forum, Oct. 10-11, 2007

2
Special Acknowledgements

• EROS
• Mike Budde
• Yingxin Gu
• Calli Jenkerson
• Ron Lietzow
• Susan Maxwell
• Shahriar Pervez
• Gail Schmidt
• Gabriel Senay

• NDMC
• Karin Callahan
• Eric Hunt
• Soren Scott
• Tsegaye Tadesse
• Brian Wardlow

Funding provided by USGS, USDA Risk Management
Agency, and NASA
3
Topics

• Status of EROS research and products
• Vegetation Drought Response Index (VegDRI)
• eMODIS
• Irrigated area mapping (MODIS)
• Normalized Difference Water Index (MODIS)
• Simple Energy Balance/Evapotranspiration (MODIS)
• MODIS Moderate Resolution Imaging
  Spectroradiometer

4
Topics

• Bigger PictureWhat is the current and future
  role of remote sensing in drought monitoring in
  the U.S.? Myths and Realities
• Disclaimer the views I express here are my own
  and in no way should
• they be construed as representing the official
  policies or opinions
• of either the US Geological Survey, the
  Department
• of the Interior, Executive Branch of the United
  States Government, any
• of the officials of these offices, of these
  entities in aggregate,
• nor of those of any other person, or
  organization. They are my
• opinions, mine alone, and I am responsible for
  them.

5
EROS Land Change System Drought Monitoring
Georegistration Compositing Surface Reflectance
Stacking Smoothing Anomaly Detection Metrics
Calculation (SOS, SG, PASG)
Existing
Intercalibration
Subscribers get Regular data over the Nation
served quickly
VegDRI models
2008
VegDRI
User/Decision Support System
Satellite Data
Remote Sensing Data Services
6
Current Year VegDRI Products
July 30, 2007
7
VegDRI Automation and Software Efficiencies

• Programming VegDRI system software
• Gained efficiencies trimmed 2 hours off of
  processing times from last year (still hampered
  by AVHRR composite processing issues)
• Server improvements stability and disk
• Testing automated system based on historical
  AVHRR input
• In Spring 2008, will test system in near-real
  time based on AVHRR and MODIS satellite inputs

8
Input Data
Data Mining and Information Discovery Subsystem
Modeling and Dissemination Subsystems
MODIS Vegetation Indices
CART analysis of input variables Iterate with
sample analyses Derive variable coefficients and
weights
Search archives
Data retrieval
Mosaic
Reprojection
Derive metrics
AVHRR NDVI
Data rescaling
Image
Extract samples
Produce Text File
Database of Model Variables
Model Parameters And Rules
Reproject
1992 National Land Cover Database (NLCD)
Calc dominant LC
Subset
Image
Model Input Data
Model Results
Extract samples
Produce Text File
Reports
State Soil Geographic (STATSGO) Database
Extract Attributes
Calculate Variables
Maps
Omernik EPA Level III Ecoregions
Rasterize
Reproject
Apply Model to Calculate Vegetation Drought
Response Index (VegDRI)
Image
Subset
Produce Text File
Extract samples
of Land in Farms in Irrigation (USDA)
Calculate variables
Polygon encoding
U.S. Drought Monitor DSS
Standardized Precipitation Index (bi-weekly)
Data retrieval
Subset
Compile database
Reproject
Segment seasons
Rescale data
Palmer Drought Severity Index (bi-weekly)
Interpolate surface
Image
Extract samples
Produce Text File
Dynamic Inputs
New input datasets
Model Formulation and Implementation Recurring
Operations
Static Inputs
Dependent variable
Processes to be automated or improved
9
Schedule for expansion
Summer 2006
Spring 2007
Winter 2007
Summer 2008
10
2007 VegDRI Map Access
http//gisdata.usgs.gov/website/Drought_Monitoring
/viewer.php
QuickViews
http//www.drought.unl.edu/vegdri/VegDRI_Main.htm
Dynamic Map Viewer
11
(No Transcript)
12

• MODIS sensor shows spectral and spatial
  resolution improvements compared to the widely
  used AVHRR sensor

The channels and the optical spectral
sensitivities for MODIS, AVHRR, and LANDSAT
13
(No Transcript)
14
VegDRI Model Improvements

• Improve quality and resolution of irrigated
  agricultue

15
VegDRI Inputs Original Irrigation Layer
Census of Agriculture statistics land cover mask
16
Prototype 2002 Irrigated Lands Map
17
Irrigated Agriculture Methodology
MODIS Annual Peak NDVI
County irrigated area statistics
Land cover mask
18
MODIS 250 meter NDVI time-series
19
Satellite Derived Vegetation Indices

• NDVI related to energy absorbed by the vegetative
  canopy to fuel photosynthesis (integrated
  measure)
• NDWI responds to changes in the water content
  (absorption of SWIR radiation) and spongy
  mesophyll in vegetation canopies

20

• NDWI MODIS sensor shows spectral and spatial
  resolution improvements compared to the widely
  used AVHRR sensor

The channels and the optical spectral
sensitivities for MODIS, AVHRR, and LANDSAT
21
(No Transcript)
22
East Amarillo Fire
23
Soil moisture sites in the OK Mesonet
24
Simplified Energy Balance Approach to Monitor
Actual Evapotranspiration
Step 1 Use MODIS thermal data to develop ET
fractions (comparable to the combination of crop
coefficient Kc and soil stress Coefficient
Ks) Step 2 Produce actual ET estimates by
multiplying the ET fractions by a global
reference ET (ETo), produced at EROS
ET ETfrac ETo
Thot Tx ETfrac
--------- Thot - Tcold
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
Future Plans for the Simple Energy Balance Work

• Collect feedback on 2006 products
• Continue to process central Great Plains data
  (2000 2006) for flash drought investigation
• Part of the U.S. Water Census?

31
Drought Monitoring at County and Sub-county Scales

• A major goal is to improve the detail of
  decision-making products
• This is stated in the 2004 Western Governors
  Association Report
• A clear way to accomplish this is to incorporate
  gridded inputs at higher resolutions, that is
  remote sensing products.

32
Bigger PictureWhat is the role of remote sensing
in drought monitoring in the U.S.? Myths and
Realities

• One goal is to improve the detail of
  decision-making products to sub-county
• This is stated in the 2004 Western Governors
  Association Report
• A clear way to accomplish this is to incorporate
  gridded inputs at higher resolutions, aka remote
  sensing.

33
Principal Drought Monitor Inputs
CPC Daily Soil Model
USGS Streamflow
Palmer Drought Index

30-day Precip.
USDA Soil Ratings
Satellite Veg Health
34
Principal Drought Monitor Inputs
CPC Daily Soil Model
USGS Streamflow
Palmer Drought Index

30-day Precip.
USDA Soil Ratings
35
Remote Sensing in Drought Monitoring Myths and
Realities

• ? Are there barriers to using remote sensing
  products in the USDM?
• ? What are they and can they be eliminated or
  minimized?

36
Remote Sensing Myths and Realities

• Related to period of record
• Related to product latency or production
  schedule
• Related to mismatch between the product design
  and the drought application

37
DRAFT USDM Requirements for Remote Sensing
Vegetation Input Data

• Consistent
• Reliable
• In anomalous vegetation signal, drought effects
  separated from non-drought
• Seasonality/phenology taken into account
• Familiar drought-like categorization
• Timely delivery (to USDM authors by Mon/Tues)
• Current information (latency not more than 48
  hoursideally within 24 hours)

38
Strengthen Partnerships

• NIDIS Knowledge Assessment Workshop (probably
  next February)
• Contributions of Satellite Remote Sensing to
  Drought Monitoring
• Planned workshop report a compendium of current
  and upcoming national remote sensing products,
  their characteristics and uncertainties

39
The remote sensing community can

• Improve response to drought community
  requirements
• Document/provide improved explanations for
  products that are tailored to the users
• Increase direct involvement with the drought
  monitoring community by requesting feedback

40
The drought community can

• Increase feedback (especially constructive
  criticism) on products
• Test and evaluate products
• Facilitate understanding that, just like drought
  indices, there is no single remote sensing
  product that will track both short-term and
  long-term drought signals

41

• Jesslyn Brown
• 605.594.6003
• jfbrown_at_usgs.gov
• Comments and Questions?