I'm a research student who's just getting started in the field of remote sensing and GIS. Therefore I kindly request your valuable support to solve my question.
I downloaded the MODIS band 1 and band 2 files using https://ladsweb.modaps.eosdis.nasa.gov/search/ in tif format.
Then I calculated the NDVI using QGIS. But my NDVI value range is -119 to 141.
Please help me to solve this matter.
How to calculate NDVI from MODIS MOD09GQ?
-
5479584.asdf
- Posts: 8
- Joined: Mon May 30, 2022 12:12 am America/New_York
How to calculate NDVI from MODIS MOD09GQ?
- Attachments
-
- raster calculate step.PNG (50.57 KiB) Not viewed yet
-
- Capture qgis.PNG (239.75 KiB) Not viewed yet
Filters:
-
LP DAAC - jwilson
- User Services
- Posts: 341
- Joined: Mon Sep 30, 2019 12:39 pm America/New_York
- Has thanked: 10 times
Re: How to calculate NDVI from MODIS MOD09GQ?
Hi @5479584.asdf
We have forwarded your inquiry to our subject matter expert for further assistance. We have not received a reply back from the PI yet on this question.
We have forwarded your inquiry to our subject matter expert for further assistance. We have not received a reply back from the PI yet on this question.
-
LP DAAC - afriesz
- Subject Matter Expert
- Posts: 71
- Joined: Tue Nov 12, 2019 4:02 pm America/New_York
- Been thanked: 3 times
Re: How to calculate NDVI from MODIS MOD09GQ?
@5479584.asdf,
I was able to replicate what you are seeing in QGIS. I don’t see a problem other than you should be quality filtering these data before you calculate the NDVI. The MODIS granule you downloaded will have a “QC_250m_1” layer in it that you should use to quality filter your data. The QA bit flags for the QC_250m_1 layer is provided in Table 8 of the User Guide (https://lpdaac.usgs.gov/documents/925/MOD09_User_Guide_V61.pdf) on page 19.
To better visualize the data you are working with, open the properties for your NDVI layer (i.e., ndvi in your screenshot), then select “Symbology”. Next…
- Change “Render type” to “Singleband pseudocolor”
- Change “Min” to “0” and “Max” to “1”
- Select a color ramp (I like BrBG for NDVI)
- Under the table, change “Mode” to “Equal Interval” and increase the “Classes” to “11”
- Finally, click “OK”
Hopefully that will give you a better view of the data.
I was able to replicate what you are seeing in QGIS. I don’t see a problem other than you should be quality filtering these data before you calculate the NDVI. The MODIS granule you downloaded will have a “QC_250m_1” layer in it that you should use to quality filter your data. The QA bit flags for the QC_250m_1 layer is provided in Table 8 of the User Guide (https://lpdaac.usgs.gov/documents/925/MOD09_User_Guide_V61.pdf) on page 19.
To better visualize the data you are working with, open the properties for your NDVI layer (i.e., ndvi in your screenshot), then select “Symbology”. Next…
- Change “Render type” to “Singleband pseudocolor”
- Change “Min” to “0” and “Max” to “1”
- Select a color ramp (I like BrBG for NDVI)
- Under the table, change “Mode” to “Equal Interval” and increase the “Classes” to “11”
- Finally, click “OK”
Hopefully that will give you a better view of the data.
-
3200937612
- Posts: 1
- Joined: Fri Nov 03, 2023 6:58 am America/New_York
Re: How to calculate NDVI from MODIS MOD09GQ?
Could you please explain how to use the QC_250 band to qualify the two reflec bands before ndvi calculation?
Thanks in advance
Thanks in advance
-
LP DAACx - lien
- Posts: 248
- Joined: Thu Jun 25, 2020 9:51 am America/New_York
- Been thanked: 9 times
Re: How to calculate NDVI from MODIS MOD09GQ?
Hello,
First, please see Table 8 in the MOD09 User Guide, this is the bit index: https://lpdaac.usgs.gov/documents/925/MOD09_User_Guide_V61.pdf
Bit 0-1 are the Modland QA bits and are probably the most important for your purpose. Pixels with a value of 00 in the 0-1 bits are considered very good quality all around and they are the ones you will want to use. Bear in mind, however, that this is a daily product so there may be a high incident of pixels that are not 00. If you decide to use pixels with a value of 01 this is where you may want to check the other Quality bits to see what is less then perfect with these pixels.
When you load the QC layer you will get pixel values in decimal and you will want to change it to binary. Lets use a "for instance". Say you have a pixel value of 4088. this converts to a binary value of 0000111111111000. These are read right to left. The Modland Quality bits are the first two on the right and they are 00 (good pixels).
You may find that the majority of pixels are not 00 because this is a daily product. This is the primary reason that the MODIS Science team uses the MOD09Q1 8-day composite to produce the MOD13Q1 Vegetation indices because it uses the best available pixel from 8 days of acquisitions so the chances are much higher to get good pixels.
Please let us know if you have any questions.
Thanks,
Brett
First, please see Table 8 in the MOD09 User Guide, this is the bit index: https://lpdaac.usgs.gov/documents/925/MOD09_User_Guide_V61.pdf
Bit 0-1 are the Modland QA bits and are probably the most important for your purpose. Pixels with a value of 00 in the 0-1 bits are considered very good quality all around and they are the ones you will want to use. Bear in mind, however, that this is a daily product so there may be a high incident of pixels that are not 00. If you decide to use pixels with a value of 01 this is where you may want to check the other Quality bits to see what is less then perfect with these pixels.
When you load the QC layer you will get pixel values in decimal and you will want to change it to binary. Lets use a "for instance". Say you have a pixel value of 4088. this converts to a binary value of 0000111111111000. These are read right to left. The Modland Quality bits are the first two on the right and they are 00 (good pixels).
You may find that the majority of pixels are not 00 because this is a daily product. This is the primary reason that the MODIS Science team uses the MOD09Q1 8-day composite to produce the MOD13Q1 Vegetation indices because it uses the best available pixel from 8 days of acquisitions so the chances are much higher to get good pixels.
Please let us know if you have any questions.
Thanks,
Brett