Remote quantification of soil composition characteristics using an integrated hyperspectral remote sensing approach

Lead Research Organisation: University of Hull
Department Name: Geography


Soils are a key component of terrestrial ecosystems with Soil Organic Carbon (SOC) being a critical part of the carbon cycle and balance, the largest terrestrial carbon pool and an important indicator of soil fertility. Even a small change in the SOC can substantially affect not only the climate but also the stability of ecosystems, because of its decisive role in the exchange of carbon between the soil and atmosphere and plant growth/production. Therefore, understanding the spatio-temporal changes of SOC is of critical importance to evaluate the feedbacks between the terrestrial C cycle, climate change and the maintenance of ecosystem functions. Soils are a complex mixture of organic and inorganic constituents with different physical and chemical properties, that show large variability at site and field scales. The capability to accurately measure SOC at high spatial and temporal resolution over large areas is therefore essential in order to inform on the effectiveness of potential sustainable agricultural practices and soil carbon sequestration approaches. Current field-based approaches provide a limited understanding of the nature, scale & spatial variability of SOC loss resulting in empirical uncertainties that may be amplified when modelling Carbon dynamics at larger scales. Hence our ability to accurately assess the impacts of climate change & farming practices in a way that will mitigate against SOC loss, is severely limited.

Remote Sensing based approaches using images covering the Visible-to-Near InfraRed (VNIR) and Short Wave InfraRed (SWIR) offer the potential to overcome these severe sampling and cost limitations and provide a low-cost, repeatable, accurate methodology that can measure SOC at site-to-landscape scales on an operational basis. There have been a large number of research projects that have identified that there is a spectral response to increasing amounts of SOC in the soil but the instruments and the methods that these studies have implemented have left a number of uncertainties including (i) how accurately can surface SOC be determined using VNIR and SWIR spectral measurements; (ii) how does the spectrally measured surface SOC composition relate the mean SOC composition of the top 30cm of the soil horizon (bulk composition); how is the spectral response of the soil affected by illumination and surface roughness variations.

This project will use a highly novel, integrated laboratory and field-based remote sensing approach using a range of hyperspectral sensors covering the VNIR and SWIR wavelengths, supported by contact spectral measurements and traditional laboratory soil analyses methods to enable these measurement uncertainties to be resolved. Two fields with contrasting soil regimes, one using organic and the other modern, commercial, farming methods to maintain soil health and SOC have been identified. An initial field-based survey using ground spectroradiometers and UAV-mounted hyperspectral cameras will acquire datasets over these two fields. The array of hyperspectral sensors (VNIR, SWIR & LiDAR) mounted on a gantry will enable ultra-high imagery to be acquired under highly controlled conditions. These datasets will enable the uncertainties stated above to be determined and a robust, accurate relationship between remotely measured SOC composition and the real surface and bulk SOC measurements to be resolved.

A large number of satellite-based hyperspectral sensors covering the VNIR and SWIR are currently being deployed. This will, for the first time, provide researchers and government agencies with the datasets they can use to derive SOC routinely, at low cost, at landscape-scales over the entire globe. The results from this project will be able to directly assist the processing of these satellite datasets in meaningful measurements of surface and bulk SOC composition which will be able to assist directly research into spatial and temporal variability in soil health.


10 25 50