A Remote Sensing Approach for Estimating Regional Scale Surface Moisture

1
A Remote Sensing Approach for Estimating Regional
Scale Surface Moisture

• Luke J. Marzen
• Associate Professor of Geography
• Auburn University
• Co-Director AlabamaView

Research funded by Alabama Water Resources
Research Institute
2
(No Transcript)
3
AmericaView Membership
4
(No Transcript)
5
(No Transcript)
6
(No Transcript)
7

• http//www.geosociety.org/meetings/06drought/facts
  heet.pdf
• Maintain and enhance hydrologic and meteorologic
  data collection capabilities and existing data
  sets, and develop new data sets needed to improve
  assessments. Automate data collection to the
  maximum practical extent, and collect data at the
  frequency and spatial scale needed to support
  model analyses and decision-making. Fully fund
  and implement the National Integrated Drought
  Information System (NIDIS) passed by Congress in
  2006.

8

• Estimated that drought costs US 6-8 billion
  annually (Wilhite, D.A. and M.D. Svodoba. 2000)

9

• Meteorological drought is usually measured by how
  far from normal precipitation has been over a
  period of time. 
• Agricultural drought occurs when soil moisture is
  insufficient to meet crops needs to produce an
  average crop.  It may occur in times of average
  precipitation depending on soil types.
• Hydrological drought refers to deficiencies in
  surface and subsurface water supplies.

10
(No Transcript)
11
Objective

• Evaluate an approach to estimate surface moisture
  conditions using remote sensing at the regional
  scale
• Scale methods down to field level

12
Vegetation and EMR

• small bodies in leaf that contains chlorophyll
• Absorbs blue and red light, reflects green and
  NIR

13

• Normalized Difference Vegetation Index
• (Red NIR)/(Red NIR)
• Values 0-1

14

• Land Surface Temperature
• Thermal RS

15

• Past research using AVHRR has exploited the
  relationship between the Normalized Vegetation
  Index and Land Surface Temperatures to evaluate
  surface moisture status (Nemani and Running,
  1989)

16
LST and NDVI relationship

• During drier periods NDVI values fall and
  vegetation canopy temperatures increase

Drier conditions
Less dry
LST - Land Surface Temperature NDVI Normalized
Difference Vegetation Index
17
Data and Methods

• -Use NDVI and LST MODIS products
• -growing season of 2000-2003
• -Evaluate ratio of NDVI/LST as an indicator of
  surface moisture
• -compare to ground-based indices

18
The MODIS Instrument Moderate Resolution Imaging
Spectroradiometer

• Global coverage - 2330 km swath
• 36 channels - 2 _at_ 250m pixels, 5 _at_ 500m, 29 _at_
  1km
• various levels of processing
• EOS Validated products
• MOD13, MOD11

19
EOS DataGateway
Land Validation Home Site
Direct to PI Websites
http//modis.gsfc.nasa.gov/cgi-bin/texis/organigra
m/weblinks
20
MOD13

• NDVI composites uses best value
• Both 250m and 1km

21
MOD11 Land Surface Temperature

• Shown to be accurate within 1 degree K
• Averaged 2 8 day composites to match NDVI

22
NDVI/LST
23
Crop Moisture Index

• Southeast Regional Climate Center
• Mean CMI was compared to the mean of NDVI/LST on
  a Climate division basis
• NDVI
• LST

24
Table 1. Pearson's Product Correlations for
Remotely sensed variables with CMI
Duration LST-CMI NDVI-CMI WSVI-CMI

April-May 2000 -0.73 0.305 0.53
June-July 2000 -0.69 0.035 0.04
Oct-00 -0.44 -0.110 -0.2
April-May 2001 -0.57 0.340 0.66
Oct-01 -0.41 -0.162 0.36
April-May 2002 -0.83 0.720 0.776
June-July 2002 -0.74 -0.560 -0.502
Oct-02 -0.66 0.259 0.77
April-May 2003 -0.44 0.085 0.45
June-July 2003 -0.57 -0.009 0.02
Oct-03 -0.55 0.228 0.43
25
Period 4 for entire southeast
26
Conclusions

• The ratio of NDVI/LST may provide an effective
  indicator of surface moisture conditions
• LST performed substantially better in our three
  year study

27
Future work

• Economic Study
• local scale/field level

28
Atlas thermal sensor 1m resolution
High crop yield red
Cool temps red not done by this geographer