I am wanting to make a graph of annual max NDVI and mean annual temperature, like Figure 2 in this paper: Raynolds, M., Magnússon, B., Metúsalemsson, S. and Magnússon, S.H. (2015) Warming, sheep and volcanoes: Land cover changes in Iceland evident in satellite NDVI trends. Remote Sensing, 7 (8). Pp.9492-9506. - I am aware that they have used GIMMS and I am using MODIS data.
My issue is that when I look at my MODIS geotiff files in ArcMap, my values are not like theirs. I can view the maximum, mean and standard deviation in the properties tab but none of these are decimals. For example, the maximum of my 2010 MODIS file is 9994, the mean is 4433.79 and the standard deviation is 2647.54. My table looks like Table 1 here: ABoVE: MODIS-derived Maximum NDVI, Northern Alaska and Yukon Territory for 2002-2017 (ornl.gov) https://daac.ornl.gov/ABOVE/guides/Alaska_Yukon_NDVI.html
How do I get these values to be decimals?
I know it is NOT simply scaling them by 0.0001 (the MODIS scale factor) because the values in Figure 4 in the link above are not the same as the values for the corresponding year in Table 1.
I am using MYD13Q1 250m
I hope that makes sense - I am extremely stuck and cannot figure it out. Please help!
I have included the tables and graphs I am referring to if you cannot access them
Martha Higgs, I'm not sure which specific MODIS product (MOD13Q1, MYD13Q1 etc.) you are using, however, if you go through the MODIS NDVI Products User Guide that is available at NASA GSFC MODIS webpage, they make it pretty clear that all 6 Vegetation Indices products are packed in INT16 Data Types, has a valid range usually between -2000 and 10000, and uses a scale factor of 0.0001. This range corresponds to -0.2 to 1 in NDVI terms. Check out the links for yourself, especially the Tables in the Users Guide. Links are attached below.
Product Guide: https://vip.arizona.edu/documents/MODIS/MODIS_VI_UsersGuide_June_2015_C6.pdf
Overview: https://modis.gsfc.nasa.gov/data/dataprod/mod13.php
Product Details: https://modis.gsfc.nasa.gov/data/dataprod/mod13.php
please refere to
https://modis.gsfc.nasa.gov/data/dataprod/mod13.php
you are most welcome for further query if any..
Thanks for answering, however I know the problem is NOT to do with applying the scale factor.
My issue here is that I have a table like the one linked above of the geotiff properties and want to get the Max NDVI as a decimal, like in Figure 2 from Raynolds, et al. (2015) and figure 4 in the link I provided.
Do you know how I convert this as the example Max NDVI graph over time and the geotiff table values in the link do not correspond.
Martha Higgs thanks for clarifying the question. I always end up on a 'Page not found' web page when I click the ORNL link that you referred to in your question. Does anybody else see the same error message?
I had to search a bit to get to the webpage you were referring to, but found it eventually. Anyway, if I understand you correctly, you are saying that Table 1. 'Properties of the GeoTIFF files' do not match with the graph plotted in Figure 4., on ORNL webpage, right?
According to what I understand from the ORNL webpage, Table 1 is referring to the NDVI Statistics (before rescaling) of the ENTIRE image for each year. So, for example, when Table 1. mentions 'Minimum Value' (2001), 'Maximum Value' (9996), 'Mean' (7459.2) and Std Deviation (991.4) for the file max_NDVI_2002.tif, it is actually referring to the entire image. There might be thousands of pixels in this image, if not millions. While the graph plotted in Figure 4. has no direct correlation with Table 1. because it is only talking about a (random) pixel from the entire image as an example. The pixel chosen is obviously not the pixel with the highest NDVI value. So, Figure 4. is NOT a graph plotted of the maximum NDVI values in the entire image. Observe carefully the description associated with Figure 4: Long-term maximum NDVI for a 250-m pixel. In this example, the long-term maximum NDVI (black marker symbol) was 0.87 from 2012 for the pixel. That's just semantics!
Let me know if this helps or it confuses you more? I would love to know the opinion of others. May be I am wrong here.
James Varghese That makes sense, thank you for explaining this!
Do you know how I get the max NDVI values like they have done in Figure 4 so I can make a similar plot as they are not plotted from the table of GeoTIFF properties? Do I select random values?
Thanks in advance!
Martha Higgs, there are so many options to create a time series plot like in Figure 4. You just need to know the sample locations from where you want to extract the NDVI values after the multi-date raster files are stacked on top of each other. This can be done easily in Google Earth Engine, ENVI, ArcGIS, Python, R, QGIS, you name it....
If you are doing this in ENVI, the idea is to first of all layer stack all the raster files (e.g., NDVI GeoTIFF files) and extract the NDVI values from the locations that you desire to study. Check this discussion page for an example https://www.researchgate.net/post/How-can-I-export-region-of-interest-ROI-for-both-X-and-Y-axis-valuex-of-two-variables-in-ENVI-53
If you are doing this in ArcGIS, you can look at the 'Extract values from multiple rasters at point locations' (Extract Multi Values to Points) option. Follow these links for more info: https://www.esri.com/arcgis-blog/products/analytics/analytics/how-to-extract-raster-values-at-point-locations/
https://desktop.arcgis.com/en/arcmap/10.3/tools/spatial-analyst-toolbox/extract-multi-values-to-points.htm
And for options to do it in Google Earth Engine there are several tutorials on YouTube https://www.youtube.com/results?search_query=how+to+create+time+series+plot+ndvi+modis+
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps