While calculating agriculture yield and production, NDVI is correlated with yield and extrapolated over the entire study area. Is NDVI a good indicator for agriculture yield and how accurate is it?
I know NDVI which is the Normalized Difference Vegetation Index is a formula employ to quantify the density of plant growth on the Earth. Using: NDVI = (NIR — VIS)/(NIR + VIS) to extrapolate agriculture yield and production may not give the accurate picture, since it may not capture total yield of most agricultural crops.
NDVI can be an indicator for yield, but it's quality as indicator for yield strongly depends of the type of crop you're aiming. You've got to keep in mind 2 points :
1. The correlation between NDVI and yield is quite satisfactory when the yield directly relies on the leafs (eg. Grass or salad, ...). It can be less relevant when the yield depends on small fruits for plants with much leafs (eg. strawberries)
2. The NDVI is not relevant for crops with 3D structure (eg. fruit tree growing or viticulture, ...). The density of vegetation can not be taken in account with NDVI.
Indeed, NDVI is mainly good to describe the vegetative expression for 2D crops.
I used NDVI for estimate the water requirements for the whole plot areas, I think NDVI also can be good indicator for the productivity, but, I don't know if there are particular model can estimate the productivity of cultivated crops based on NDVI
I used NDVI to estimate the amount of Nitrogen needed to top dress Hard Red Winter Wheat just before it joints in Northern Oklahoma the 3 or 4 years I wrote code for the prototypes built by Biosystems & Ag Engineering at Oklahoma State. Reflected Solar NDVI was an excellent method as if the proper wavelengths were used. Averging 128 NDVI readings or more per meter proved extremely robust even with poor signal to noise ratios.
Others in the department found NDVI values between various wavelengths effective for fertilizing other crops. The same equipment loaded with different chemicals scanning fallow fields reduced the herbicide by 50% need to control weeds. I didn't explore the importance of wavelength on differentiating plants from the soil. The project wasn't at the stage of weed identification. The algorithm never got past trying to kill it if it was green. All the effort was put in to fertilizing wheat.
The USDA used NDVI from the first MODIS Satellites effectively in the early 1970's to identify field crops.
Selecting the best wavelengths for the desired results is critical to obtaining the best NDVI results from the available data. I haven't kept up, so I expect there are still gaps in the spectral coverage camera's on the MODIS satellites.
An often overlooked source of error are minerals that are reflective to the wavelengths used to calculate NDVI. The time of day needs correction before 9:00 A.M. and after 3:00 P.M. Solar time. If it is not already corrected.
I tried measuring NDVI up to 400x with a microscope. Although the results are incomplete they look promising.
Using combination of NDVI and Climatic data and calculating ET would be more accurate as both aspects will be covered i.e. crop health and water supply
- Badges
- Science topic
- Similar topics
- Agriculture