A Remote Sensing Approach to Characterising Soil Moisture Regimes on Poorly Drained Soils in Ireland

Abstract:

Soil drainage capacity is the degree and frequency at which the soil is free of saturation. It influences land use and management, soil nutrient cycling and greenhouse gas fluxes; accurate information on soil moisture and drainage conditions is crucial for crop production and management and fundamental in developing strategies to adhere to environmental sustainability goals.

Temporal studies of soil moisture act as a proxy for drainage capacity which is particularly important in Ireland where approximately 50% of the soils are classified as “marginal” which is limited mainly by poor drainage status. These poorly drained soils negatively affect plant growth and productivity. Timely and accurate information on soil moisture would allow for precision management strategies and aid in designing effective interventions on farms with regard to artificial drainage works.

The design of agricultural drainage systems on a given site is dictated by the soil characteristics in the near surface (<2m depth) and outfall conditions, which are assessed by information on soil moisture, soil type and hydrology. Such data is conventionally acquired by in-situ point sampling techniques which are costly and time consuming. Remote sensing has the potential to provide a solution by allowing simultaneous coverage of large geographic areas, quickly and in a cost effective manner. This project aims to use optical remote sensing data from Sentinel 2 to derive information on soil moisture conditions on select sites in Ireland. We explore the use of the additional red edge band present in Sentinel 2 to model soil moisture . Remote sensing derived vegetation indices such as the Normalised Difference Vegetation index (NDVI), Enhanced Vegetation Index (EVI) and Normalised Difference Red Edge Index (NDRE) for the years 2015-2020 have been non-linearly modelled with Short wave Transformed infrared Reflectance (STR). We compare the effectiveness and similarities of different vegetation indices in modelling soil moisture conditions over our study area. We aim to identify saturated and non-saturated areas and to produce soil moisture maps. We will show that when compared to previously used linear models, the non-linear model is novel and better suited for Ireland; which is dominated by wet conditions for most of the year.