Comparison of the MARACOOS AVHRR Sea Surface Temperature product to NDBC sites Matthew Oliver, Matt Shatley, Taylor Daley, Mike Crowley University of Delaware Rutgers University
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Comparison of the MARACOOS AVHRR Sea Surface Temperature product to NDBC sites
Matthew Oliver, Matt Shatley, Taylor Daley, Mike CrowleyUniversity of DelawareRutgers University
MARACOOS - International Constellation of Satellites – Since 1992X-Band (installed 2003)
L-Band (installed 1992)Sea Surface Temperature - SST
Ocean Color<River PlumesOcean Blooms>
Combined SST & Color Water Mass Boundaries
The Problem with AVHRR - Clouds
Both temperature and edge effects
Testing for cloud contamination (Three tests)
Test 1: Time Test: A measurement fails the time test if it is more than 2 degrees cooler than the mean of the satellite passes in the previous 72 hrs.
Test 2: Climatology Test: A measurement fails the climatology test if the temperature is more than 5 degrees colder than the climatologies calculated by NSIPP AVHRR Pathfinder and Erosion Global 9km SST Climatology (Casey, Cornillon) A description of this climatology can be found at ftp://podaac.jpl.nasa.gov/pub/documents/dataset_docs/nsipp_climatology.htm
Test 3: Sequential Test: In this test, the previous 72 hours or satellite passes are first passed through a climatology test, then passed through the time test. It is a sequential test, rather than two independent tests.
If any measurement is fails one of these three tests, all neighboring measurements within 3km are also automatically failed.
Temporally and Spatially interpolated climatology
Coastal Pixels are nearest-neighbor interpolated
Coding of the Tests
0: Passed all tests1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
3: Failed test 1 and 25: Failed test 1 and 46: Failed test 2 and 47: Failed test 1, 2, and 4
Pixels that failed time test1, 3, 5, 7Pixels that failed clim test2, 3, 6, 7Pixels that failed seq. test4, 5, 6, 7
Data compression
The Problem with AVHRR - Clouds
Both temperature and edge effects
Eliminating Flag >= 7
1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
It had to fail every test
Eliminating Flag >= 6
2: Failed the clim. test4: Failed the seq. test.
Or1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
Eliminating Flag >= 5
1: Failed the time. test4: Failed the seq. test.
Or2: Failed the clim. test4: Failed the seq. test.Or1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
Eliminating Flag >= 4
4: Failed the seq. test.
Or1: Failed the time. test4: Failed the seq. test.Or2: Failed the clim. test4: Failed the seq. test.Or1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
Eliminating Flag >= 3
1: Failed time test2: Failed clim test
Or4: Failed the seq. test.Or1: Failed the time. test4: Failed the seq. test.Or2: Failed the clim. test4: Failed the seq. test.Or1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
Eliminating Flag >= 2
2: Failed clim testOr (3)1: Failed time test2: Failed clim testOr (4)4: Failed the seq. test.Or (5)1: Failed the time. test4: Failed the seq. test.Or (6)2: Failed the clim. test4: Failed the seq. test.Or (7)1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
Eliminating Flag >= 11: Failed time testOr (2)2: Failed clim testOr (3)1: Failed time test2: Failed clim testOr (4)4: Failed the seq. test.Or (5)1: Failed the time. test4: Failed the seq. test.Or (6)2: Failed the clim. test4: Failed the seq. test.Or (7)1: Failed the time test2: Failed the clim. test4: Failed the seq. test.
How does this compare to available platforms and NDBC bouys?
SST data was downloaded for 205 NDBC sites from http://coastwatch.pfeg.noaa.gov/erddap/tabledap/ndbcSosWTemp.subset
(Jan 1, 2006 – September 14, 2014)
The data on this site is “initially quality controlled.”
We further quality controlled the data by visually inspecting each data set.
The downloaded buoy data set was hourly, but we interpolated each SST measurement to the satellite pass time
How does this compare to all platforms?
How does this compare to NDBC Buoys?
What does this look like with buoy records?
Unfiltered SST (red) compared to Buoy (black)
What about errors for other time integrations?
What about errors for other time integrations?
How does this compare to MUR?
Multi-scale Ultra-high Resolution (MUR) Sea Surface Temperature (SST) AnalysisGap-filled, ~1km res, 24 hr delay.
https://podaac.jpl.nasa.gov/Multi-scale_Ultra-high_Resolution_MUR-SSTMARACOOS AVHRR 1-day MUR 1-day
How does this compare to MUR?
MUR b
ette
r
MAR. A
VHRR bet
ter
MUR b
ette
r
MAR. A
VHRR bet
ter
How does this compare to MUR?
How does this compare to MUR?
Conclusions
MARACOOS AHVRR de-clouded products1. Available at 1-km in real time at the pass, 1, 3, 7
day aggregations2. Individual passes are flagged for clouds3. Individual users can opt to use different
combinations of the tests4. Comparisons to in situ records vary, but RMS
errors to NDBC buoys ~0.7 deg C.
JPL MUR 1km product5. Available at 1-km at 1day aggregation with 24
hour delay6. Data are gap filled with multiple platforms7. Does not appear to preserve all hi-resolution
features8. Comparisons to in situ records vary, but RMS
errors to NDBC buoys ~0.5 deg C.9. More variation in bias across platforms
MARACOOS Cloud Filtered SST Resources
http://tds.maracoos.org/thredds/SST.htmlhttp://tds.maracoos.org/thredds/MODIS.htmlhttp://modata.ceoe.udel.edu/web_kmzs/
Soon to include MUR!!We are working with JPL for a MARACOOS subset that will be served via this portal for the MABPOM communityAsk me for the error tables.
Related Documents
