Digital Soil Mapping explained in detail a publication under review. Briefly, Maps of soil condition were created for the threats of acidification using pH as indicator. Digital soil maps of pH were created applying the scorpan approach (McBratney et al., 2003) fitting quantile regression forest models (Meinshausen & Ridgeway, 2006) for predicting the soil indicators for both depth intervals (0-20 cm, 20-40 cm). Croplands, water and artificial areas were masked because we did not have pH observations from these land uses in the calibration dataset. The thresholds of soil condition were defined by soil monitoring unit (see map of soil monitoring units for the Basque Country (Román Dobarco et al. 202410.5281/zenodo.14674358), using data from semi-natural forest of native species (e.g., beech, oaks, helm oak, etc.) as reference soils. The thresholds were defined with different percentiles (5%, 12.5% and 25%) from the distribution of reference soils. If the pH of the forest plantations was lower than the threshold, then the soil was considered in “poor condition” (coded with the value 0), and if it was equal or greater, in “good condition” (coded with the value 1). The assessment was applied at each pixel.

The environmental covariates included: Relief covariates were derived from a digital elevation model (DEM) at 25 m resolution (Eusko Jaurlaritza / Gobierno Vasco, 2016) using GRASS GIS 8.3 (GRASS Development Team, 2024) or SAGA GIS (9.2.0) (Conrad et al., 2015): elevation, slope (%), northerness, easterness (cosine and sine transformation of aspect respectively) and interaction between northerness and slope informing of topographic exposure and microclimate, SAGA topographic wetness index (TWI) as a proxy for soil moisture, and valley depth and standardised height as relative elevation metrics, profile and tangential curvature; Climate covariates were the mean annual air temperature (°C), temperature seasonality (standard deviation of monthly temperatures), maximum temperature of the warmest month (°C), mean annual precipitation (mm), and precipitation seasonality (coefficient of variation). Climatic variables were downloaded from CHELSA dataset version 2.1 (Brun et al., 2022); Parent material was represented with lithology and regolith maps at scale 1:25,000 for the Basque Country (Eusko Jaurlaritza / Gobierno Vasco, 1999) available in vectorial format. There were 20 lithology classes that were grouped into 7 broad classes. The regolith had 5 thickness categories (< 0.5 m, 0.5 m – 1 m, 1 m – 2 m, 2 m – 4 m, and > 4 m); Organisms (vegetation) was characterised with the median normalised difference vegetation index (NDVI) for the four quarters of the year, and annual standard deviation as proxies for vegetation. NDVI layers were processed with Google Earth Engine with time series of the Landsat Collection 8 Level 2 Tier 1 (2014-2020) (Crawford et al., 2023; USGS, 2024) after masking clouds and shadows, excluding images with more than 15% of land covered by clouds, and averaged by quarter (January-March, April-June, July-September, October-December). The quantile regression forest models were evaluated with 10-fold cross-validation using root mean squared error (RMSE), bias or mean error (ME), coefficient of determination (R²), and Lin’s concordance correlation coefficient (CCC) (Lin, 1989) as validation statistics. Uncertainty estimates for soil condition assessment were calculated by mapping the probability that the threshold was surpassed from the full-conditional probability distributions of the predicted indicators and expressed as percentage.

The QRF models for pH had good performance in cross-validation, with RMSE = 0.79, ME = 0.01, R² = 0.53 and Lin’s CCC = 0.67 for 0-20 cm. Similarly, the model for 20-40 cm obtained a RMSE = 0.81, ME = 0.02, R² = 0.53 and Lin’s CCC = 0.67. The maps of soil pH reflected the influence of lithology and climatic variables, with annual precipitation, temperature seasonality, and maximum temperature of the warmest month as the three most important variables for both depths, followed by lithology and NDVI Q3 for 0-20 cm, and elevation and annual mean temperature for 20-40 cm. The pH was lower in the Atlantic basin following the precipitation gradient, and higher in areas dominated by calcareous rocks and superficial deposits. The surface of forest plantations in the Basque Country considered in unhealthy condition, estimated from pH maps, ranged between 1.6% and 42%.

-      Date of publication: 31/01/2025

-      Reference period: 2024

-      Date of creation: 19 July ‎2024

-      Last modification: 19 July ‎2024

-      Type of data: GeoTIFF file

-      Purpose of the data: Soil condition assessment using soil pH as indicator.

-      Lineage: First version. The sources of the input data will be described in the accompanied peer review publication and Deliverable 2.1.

-      Resolution: 25 m

-      Location details (geographic extent):

o   North: 4,811,750 m

o   South: 4,700,850 m

o   West: 461,050 m

o   East: 606,500 m

-      Projected Coordinate system: ETRS 1989 UTM Zone 30N (EPSG: 25830)

-      Geographic Coordinate system: ETRS 1989 (EPSG: 4258)

-      Creator or author of the data: Mercedes Román Dobarco, Alex McBratney, Sophie Cornu, Jorge Curiel Yuste.

-      Contact: Mercedes Román Dobarco (mercetadzio@gmail.com )

-      Access and licensing information: These maps are available under the CC-BY 4.0 License.

-      DOI: 10.5281/zenodo.14700581

-      Associated publications / Conference presentation: Román Dobarco, M., McBratney, A., Cornu, S., Curiel Yuste, J. Monitoring the condition of forest soils in the Basque Country (Spain). An application of pedogenon mapping for policy implementation. Presented at the Centennial of the IUSS (May 19-21, 2024), Florence, Italy.

-      Dataset citation: Román Dobarco, M., McBratney, A., Cornu, S., Curiel Yuste, J. 2025 Digital soil maps for assessing the condition of forest soils of the Basque Country using pH as indicator. DOI: 10.5281/zenodo.14700581

-      File size: 3.2 MB – 34.7 MB.

-    Keywords: Pedogenon, digital soil mapping, soil monitoring, Soil Monitoring and Resilience Law, soil security.

References

Brun, P., Zimmermann, N.E., Hari, C., Pellissier, L., Karger, D.N., 2022. CHELSA-BIOCLIM+ A novel set of global climate-related predictors at kilometre-resolution. http://dx.doi.org/10.16904/envidat.332

Crawford, C.J., Roy, D.P., Arab, S., Barnes, C., Vermote, E., Hulley, G., Gerace, A., Choate, M., Engebretson, C., Micijevic, E., Schmidt, G., Anderson, C., Anderson, M., Bouchard, M., Cook, B., Dittmeier, R., Howard, D., Jenkerson, C., Kim, M., Kleyians, T., Maiersperger, T., Mueller, C., Neigh, C., Owen, L., Page, B., Pahlevan, N., Rengarajan, R., Roger, J.-C., Sayler, K., Scaramuzza, P., Skakun, S., Yan, L., Zhang, H.K., Zhu, Z., Zahn, S., 2023. The 50-year Landsat collection 2 archive. Science of Remote Sensing 8, 100103. https://doi.org/10.1016/j.srs.2023.100103

Conrad, O., Bechtel, B., Bock, M., Dietrich, H., Fischer, E., Gerlitz, L., Wehberg, J., Wichmann, V., Böhner, J., 2015. System for Automated Geoscientific Analyses (SAGA) v. 2.1.4. Geosci. Model Dev. 8, 1991–2007. https://doi.org/10.5194/gmd-8-1991-2015

European Commission, 2023. Proposal for a DIRECTIVE of the EUROPEAN PARLIAMENT and of the COUNCIL on Soil Monitoring and Resilience (Soil Monitoring Law) (No. COM(2023) 416 final). European Commission, Brussels.

Eusko Jaurlaritza / Gobierno Vasco, 2016. Modelo Digital del Terreno (MDT) de 1m de la Comunidad Autónoma del País Vasco. Año 2016.

Eusko Jaurlaritza / Gobierno Vasco, 2019. Cartografía de litología y permeabilidad de la Comunidad Autónoma del País Vasco a escala 1:25.000.

General Secretariat of the Council, 2024. General approach on the Proposal for a DIRECTIVE of the EUROPEAN PARLIAMENT and of the COUNCIL on Soil Monitoring and Resilience (Soil Monitoring Law) (No. 2023/0232(COD)). Council of the European Union, Brussels.

GRASS Development Team, Landa, M., Neteler, M., Metz, M., Petrášová, A., Petráš, V., Clements, G., Zigo, T., Larsson, N., Kladivová, L., Haedrich, C., Blumentrath, S., Andreo, V., Cho, H., Gebbert, S., Nartišs, M., Kudrnovsky, H., Delucchi, L., Zambelli, P., Lennert, M., Mitášová, H., Chemin, Y., Pešek, O., Barton, M., Tawalika, C., Ovsienko, D., Bowman, H., 2024. GRASS GIS. https://doi.org/10.5281/zenodo.10817962

Lin, L.I.-K., 1989. A Concordance Correlation Coefficient to Evaluate Reproducibility. Biometrics 45, 255. https://doi.org/10.2307/2532051

McBratney, A.B., Mendonça Santos, M.L., Minasny, B., 2003. On digital soil mapping. Geoderma 117, 3–52. https://doi.org/10.1016/S0016-7061(03)00223-4

Meinshausen, N., Ridgeway, G., 2006. Quantile regression forests. Journal of machine learning research 7.

USGS, 2024. Landsat 8 Collection 2 Level-2 image courtesy of the U.S. Geological Survey.