MIR spectroscopy combined with meteorological data can estimate soil compaction risks in top and subsoils.

Soil compaction is an important physical characteristic that affects agricultural productivity by increasing soil density, which reduces the volume of a given soil mass. Due to the higher compaction, plant roots find resistance in penetrating deeply into the soil, limiting their access to essential nutrients and moisture, impacting the plant health with lower levels of N, P and K, resulting in lower productivity. Soil compaction can also reduce soil porosity, aeration, carbon mineralisation/sequestration and increasing the production of greenhouse gases through denitrification in anaerobic sites. Besides that, soil compaction can cause surface runoff and erosion, increasing the risk of flooding and soil loss. A partial recuperation of compacted soils is an expensive and labour-intensive task. In addition, agricultural land expansion for crops is limited. Hence, mapping agricultural areas at risk of soil compaction is essential to implement strategies to mitigate the adverse effects of soil compaction.

Soil particle size and soil drainage were used to classify topsoil's (T) compaction risk class. For subsoil (S) soils (after horizon A), the subsoil particle size, packing density (bulk density + 0.009 * clay (%)), soil drainage and field capacity days were used to estimate the compaction risks. The main problem of this strategy is that these analyses are expensive and time-consuming, i.e., soil particle size analysis requires an average time of 1 month per 100 samples and costs ~ 40.00 per sample. Bulk density analysis costs ~ € 7.00 per sample and is also time-consuming; consequently, bulk density values are mainly predicted using pedo-transfer functions in mapping studies.

To speed up the analysis and minimise the costs, vibrational spectroscopy combined with chemometrics was used to determine soil particle size and bulk density. Both parameters were combined with field capacity days (obtained from 104 national wide meteorological stations) and drainage class (obtained from Irish - Environmental Protection Agency) to map soil compaction risk areas in the northern half of the Republic of Ireland with a resolution of 4 km² (2x2km) and 1 km² grid for regional and periurban regions, respectively (Tellus achieve). To confidentially map these regions, spectral control charts based on PCA were used to identify unrepresentative sample spectra based on the spectral models used. Only samples classified as representative were predicted by the spectral models. Using this strategy, we could predict ~ 90% (T) and ~66% (S) compaction risks in non-peat soils. The prediction results showed that ~33% (T) and ~43% (S) were classified as high risks of compaction, ~19% (T) and ~23% (S) as moderate, and ~37% (T) and <1% (S) as low risks or other classes.