I wish to develop a tool to monitor soil health of a perennial tree cropping system. This data set will include some physical, chemical and biological properties of soil. We hope to ensure that this tool contains as many soil properties as possible. Can anyone offer suggestions/ information on appropriate cost- effective remote sensing techniques that could be used for this purpose?
I think that you need to decide the parameters (physical, chemical, and biological). You also need to give weights to each parameter, and definitly you need to categorize those parameters. Cooperation with specialists in agriculture is important.
I agree with Dat Dinh that you can use Radar to extract soil properties and NDVI to monitor the health of tree. but it still not a cost efective since you will need to pay for the Radar data.
on the other hand, In following article you can find one of the steps you will need to do this analysis. where you will need to measure weighs for these soil properties since not all of them has the same effect on the soil health.
http://www.sciencedirect.com/science/article/pii/S0169204608001333
For total utilization of remote sensing techniques to calculate soil health "capability" you need 1) Radar images to define the physical characteristics of the soil, 2) Hyper-spectral optical images to define the chemical and biological indices. 3) For extracting slopes and drainage network any optical satellite with stereo capability will be fine
This is feasible. I think you need to think about radiance/reflectance characteristics and wavelength bands that could be calibrated with soil properties, such as organic carbon. Normally soil with low soil carbon have reflectance in the 1-2.5 micro meter (NIR). See Geoderma 143 (2008) 73–84; Geoderma 133, 281–294; Soil Science Society America Journal 66, 988–998.
Many thanks all. The area under question is about 1500 kmsq . The validation of the remote sensing techniques will be validated with 200 soil sample based to grid sample system. I will like to know whether the sample technique is adequate for such area of coverage. Do we have other soil sampling techniques that could be used to provide a representative sample of such landscape?
I disagree with Chadi that you need hyperspectral data. You will need multi-spectral data that contains the necessary bands. You can do a literature search for the crop and soil types to help determine those bands (although Alex did provide some insight).
As for your sampling grid, do you already know the soil types and their variability in your study area? In the USA we have the USDA (not USGS as I originally stated) soil survey that we can use to determine the variability of our study areas, does your country have a similar system? If so, then you can use that map to build a representative sample of your soil types. A grid system is a reasonable approach for an unknown area, but it is likely you can miss some soil types, especially since you have a large area and relatively few samples.
Alternatively, if you do not have previous knowledge of your soil types (e.g. USDA soil survey) you may be able to use legacy Landsat or MODIS data to determine soil quality. Areas that are consistently 'greener' using a VI (NDVI, Simple Ratio, Chlorophyll Index) will likely have different soil types (and hydrological regimes, but these are also related to soil types). Creating a map of this 'greenness' may provide an alternative to a soil map to give you a sense of the variability in your study area (although the accuracy will be quite poor).
Anthony, its USDA-NRCS not USGS, and Alex, the remote sensing for organic matter, clay content, and even soil test P have been done, but the work better when you have unencumbered access to bare soil. That can be a problem.
Michael, thanks for the clarification. Yes it is the USDA for the soil survey (http://websoilsurvey.nrcs.usda.gov/app/HomePage.htm). USGS provides a wealth of many other maps/GIS data (http://www.usgs.gov/pubprod/).
Dat Dinh Ngoc, please make further clarification. Many thanks for your insightful contributions. To all, the idea is to build a "soil test technique" that can be used by farmers to monitor the health of their farms at small scale. We are optimistic that the technique will reduce the huge cost involved in laboratory soil analysis.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing