an introductory method for evaluating the modis cloud mask over water
DESCRIPTION
An Introductory Method For Evaluating the MODIS Cloud Mask Over Water. Corey Calvert UW/CIMSS, Madison, Wisconsin Andrew Heidinger NOAA/NESDIS Office of Research and Applications November 1, 2006. Outline. Method Data used Cloud mask error estimation Examples Results / Comparisons - PowerPoint PPT PresentationTRANSCRIPT
An Introductory Method For An Introductory Method For Evaluating the MODIS Cloud Evaluating the MODIS Cloud
Mask Over WaterMask Over WaterCorey CalvertCorey Calvert
UW/CIMSS, Madison, WisconsinUW/CIMSS, Madison, Wisconsin
Andrew HeidingerAndrew HeidingerNOAA/NESDIS Office of Research and NOAA/NESDIS Office of Research and
ApplicationsApplications
November 1, 2006November 1, 2006
OutlineOutline MethodMethod
Data usedData used
Cloud mask error estimationCloud mask error estimation
ExamplesExamples
Results / ComparisonsResults / Comparisons
Future workFuture work
Motivation?Motivation? Currently there are little or no globally Currently there are little or no globally
quantitative tools to evaluate the quantitative tools to evaluate the MODIS cloud mask over waterMODIS cloud mask over water
Evaluation/validation tools will be Evaluation/validation tools will be needed for the instruments aboard needed for the instruments aboard future satellites (e.g., GOES-R, future satellites (e.g., GOES-R, NPOESS VIIRS)NPOESS VIIRS)
MethodologyMethodology MW SST measurements relatively MW SST measurements relatively
insensitive to cloud and aerosol effectsinsensitive to cloud and aerosol effects
IR SST measurements will likely have a IR SST measurements will likely have a cool bias in the presence of cloudscool bias in the presence of clouds
Differences between IR and MW Differences between IR and MW measurements should be indicative of measurements should be indicative of cloudsclouds
DataData OI MW SST product combining TMI OI MW SST product combining TMI
and AMSR-E data and AMSR-E data (daily – 25 km) (daily – 25 km) ((www.ssmi.comwww.ssmi.com))
MOD28 - MODIS SST (day) and MOD28 - MODIS SST (day) and SST4 (night) products (1km)SST4 (night) products (1km)
MOD35 - MODIS cloud mask (1 MOD35 - MODIS cloud mask (1 km)km)
IR-MW Differences & Cloud IR-MW Differences & Cloud MaskMask
conf. cloud prob. cloud prob. clear conf. clear
Example Distribution of Example Distribution of Temperature DifferencesTemperature Differences
Calculating Error for Clear Calculating Error for Clear CategoriesCategories
Differences in Differences in region B are region B are assumed to be assumed to be attributed to non-attributed to non-cloud issues cloud issues
totalBAerror
A B
Calculating Error for Cloudy Calculating Error for Cloudy CategoriesCategories
C totalCerror
Points in region C Points in region C may be clear pixels may be clear pixels misidentified as misidentified as cloudy cloudy
AQUA & TERRAAQUA & TERRA
Latitudinal ComparisonLatitudinal Comparison
Day/Night Latitudinal Day/Night Latitudinal ComparisonComparison
Graphical Comparison Graphical Comparison (April 18, 2006)(April 18, 2006)
AQUA (daytime)
-10.00
0.00
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
conf. clear prob. clear prob. cloudy conf. cloudy
global error %
tropics error %
midlat error %
polar error %
global data %
tropics data %
midlat data %
polar data %
AQUA (nighttime)
-10.00
0.00
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
conf. clear prob. clear prob. cloudy conf. cloudy
global error %
tropics error %
midlat error %
polar error %
global data %
tropics data %
midlat data %
polar data %
TERRA (daytime)
-10.00
0.00
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
conf. clear prob. clear prob. cloudy conf. cloudy
global error %
tropics error %
midlat error %
polar error %
global data %
tropics data %
midlat data %
polar data %
TERRA (nighttime)
-10.00
0.00
10.00
20.00
30.00
40.00
50.00
60.00
70.00
80.00
90.00
100.00
conf. clear prob. clear prob. cloudy conf. cloudy
global error %
tropics error %
midlat error %
polar error %
global data %
tropics data %
midlat data %
polar data %
SummarySummary The MODIS cloud mask appears The MODIS cloud mask appears
consistentconsistent Between both Aqua and TerraBetween both Aqua and Terra Between day and nightBetween day and night
Latitudinal differences are seenLatitudinal differences are seen Smaller % of clear pixels and larger % of cloudy Smaller % of clear pixels and larger % of cloudy
pixels towards the polespixels towards the poles confident cloudy errors decreases towards the confident cloudy errors decreases towards the
polespoles
Future WorkFuture Work Comparison between collect 4/5Comparison between collect 4/5
Seasonal comparisonsSeasonal comparisons
Comparison between different Comparison between different instruments (e.g., AVHRR)instruments (e.g., AVHRR)
Operational Use of the MODIS Operational Use of the MODIS Cloud MaskCloud Mask
A regional real-time SST analysis is being created A regional real-time SST analysis is being created around Florida (Florida Tech)around Florida (Florida Tech) Produced 4 times per day Produced 4 times per day Combines GOES SST composites and level 2 MODIS SSTs Combines GOES SST composites and level 2 MODIS SSTs
(DB UW) via 2DVar method(DB UW) via 2DVar method Will eventually be used as lower boundary initialization Will eventually be used as lower boundary initialization
over water for ARPS/ADAS or WRF models at NWS offices over water for ARPS/ADAS or WRF models at NWS offices in FLin FL
A consistent and reliable cloud mask for the A consistent and reliable cloud mask for the MODIS SSTs is crucial to maintain the quality of MODIS SSTs is crucial to maintain the quality of the overall analysisthe overall analysis
QUESTIONS?QUESTIONS?
Acknowledgements:Acknowledgements:Richard FreyRichard Frey
Kathy StrabalaKathy Strabala