Hello,
I am working with L2A EMIT data from LP.DAAC (EMITL2ARFL). I was wondering if there is a way to calculate remote sensing reflectance (rrs) from the surface reflectance in this product?
I did a simplified conversion (rrs = surface reflectance / pi) and got values that were still quite high (see comparison attached.) I was hoping to try the more accurate rrs conversion with downwelling irradience (Ed) and solar zenith angle (θs) but could not find this info in the EMIT RFL files.
Do you have any advice?
Calculating Rrs from EMIT L2A RFL
-
LP DAAC - dgolon
- User Services
- Posts: 296
- Joined: Mon Sep 30, 2019 10:00 am America/New_York
- Has thanked: 18 times
- Been thanked: 2 times
- Contact:
Re: Calculating Rrs from EMIT L2A RFL
Hi @slsmit34 Let me check with our science team. We'll write back soon. Thanks - Danielle
Subscribe to the LP DAAC listserv by sending a blank email to lpdaac-join@lists.nasa.gov.
Sign up for the Landsat listserv to receive the most up to date information about Landsat data: https://public.govdelivery.com/accounts/USDOIGS/subscriber/new#tab1.
Sign up for the Landsat listserv to receive the most up to date information about Landsat data: https://public.govdelivery.com/accounts/USDOIGS/subscriber/new#tab1.
Re: Calculating Rrs from EMIT L2A RFL
Hi @slsmit34! I'm supporting some of the EMIT aquatic processing and wanted to ask if you could provide the image ID you're working with and what type of pixel extraction you're doing (e.g., closest pixel to AERONET-OC coordinate, 3x3 mean or median, etc.). In addition, did you enable the L2 masks before calculating Rrs?
Re: Calculating Rrs from EMIT L2A RFL
Hi m11keluis, thanks for reaching out.
Image IDs:
- 'EMIT_L2A_RFL_001_20230603T172005_2315412_008.nc',
- 'EMIT_L2A_RFL_001_20230607T154124_2315810_001.nc',
- 'EMIT_L2A_RFL_001_20230814T192838_2322613_005.nc',
- 'EMIT_L2A_RFL_001_20230818T175401_2323012_005.nc',
- 'EMIT_L2A_RFL_001_20230822T161909_2323411_003.nc'
I am doing a 3x3 mean pixel extraction at the AERONET-OC site (lat: 39.12351, lon: -76.3489). The processing steps I did prior to getting this pixel window were:
ds_geo = emit_xarray(fp, ortho=True) #Orthorectify
ds_geo['reflectance'].data[:,:,ds_geo['good_wavelengths'].data==0] = np.nan #Remove bad wavelength data
ds = emit_aqua.mask_aqua(ds_geo) #Enable masks from emit_aqua module
ds['Rrs'] = emit_aqua.rho2Rrs(ds) #Calculate Rrs using rhos/pi from emit_aqua module
I've attached another plot with all 5 of the comparisons to this post. The AERONET matches for all 5 are the points at the bottom left (They end at ~800nm). Hopefully these answer your questions!
Best,
Sami
Image IDs:
- 'EMIT_L2A_RFL_001_20230603T172005_2315412_008.nc',
- 'EMIT_L2A_RFL_001_20230607T154124_2315810_001.nc',
- 'EMIT_L2A_RFL_001_20230814T192838_2322613_005.nc',
- 'EMIT_L2A_RFL_001_20230818T175401_2323012_005.nc',
- 'EMIT_L2A_RFL_001_20230822T161909_2323411_003.nc'
I am doing a 3x3 mean pixel extraction at the AERONET-OC site (lat: 39.12351, lon: -76.3489). The processing steps I did prior to getting this pixel window were:
ds_geo = emit_xarray(fp, ortho=True) #Orthorectify
ds_geo['reflectance'].data[:,:,ds_geo['good_wavelengths'].data==0] = np.nan #Remove bad wavelength data
ds = emit_aqua.mask_aqua(ds_geo) #Enable masks from emit_aqua module
ds['Rrs'] = emit_aqua.rho2Rrs(ds) #Calculate Rrs using rhos/pi from emit_aqua module
I've attached another plot with all 5 of the comparisons to this post. The AERONET matches for all 5 are the points at the bottom left (They end at ~800nm). Hopefully these answer your questions!
Best,
Sami
- Attachments
-
- AERONET in-situ matchups are the points at the bottom left that end at ~800nm.
- EMIT_vs_Aeronet_5_cases.PNG (113.58 KiB) Not viewed yet
Re: Calculating Rrs from EMIT L2A RFL
Hi @slsmit34 !
From your plots, it looks to me like most of the comparisons (except the green line?) might have cloud or cirrus contamination. Have you looked at the RGB images to see if they look cloudy? Sometimes even clear-looking scenes can have cirrus; this shows up in the radiance data as a nonzero signal around 1390 nm. I hesitate to say for sure without looking at the full images and radiance data, but that's my best guess based on the spectra.
Another thing to note is that Rrs measured in situ often excludes sunglint, while the EMIT reflectance data currently does not. This will require a post hoc sunglint correction. We are working on refactoring our L2A code to enable sunglint modeling and removal.
David
From your plots, it looks to me like most of the comparisons (except the green line?) might have cloud or cirrus contamination. Have you looked at the RGB images to see if they look cloudy? Sometimes even clear-looking scenes can have cirrus; this shows up in the radiance data as a nonzero signal around 1390 nm. I hesitate to say for sure without looking at the full images and radiance data, but that's my best guess based on the spectra.
Another thing to note is that Rrs measured in situ often excludes sunglint, while the EMIT reflectance data currently does not. This will require a post hoc sunglint correction. We are working on refactoring our L2A code to enable sunglint modeling and removal.
David