Spatio-temporal mapping of soil water storage in a semi-arid landscape of northern Ghana – A multi-tasked ensemble machine-learning approach

Data on spatial and temporal modelling of soil water storage in the Guinea savannah zone of Northern Ghana

“…One of the objectives of work reported in the associated paper of this data article, Nketia et al. [1] was to estimate SM from Sentinel-1 data. Thus, the SM measurements were timed to coincide with the overpass of the Sentinel-1 satellite at a temporal resolution of 12 days for ten time-steps covering the whole dry season (i.e., February–June).…”

“…In a first step, in situ SM measurements were vertically discretized into six depth intervals (i.e., 0–5, 5–15, 15–30, 30–40, 40–60, and 60–100 cm) following the GlobalSoilMap specifications [2] . In a second step, SWS at each data point was recursively profiled as a function of the measured in situ SM, bulk density and the effective soil thickness between two soil layers [1] . By this approach, we accounted for the differential availability of SWS critical to the management of shallow and deep-rooted plants notable to the study area.…”

“… The data provides information on space-time SM and SWS over 0–100 cm soil depth for key agricultural benchmark soils of the Guinea savannah zone of Ghana. It provides useable information on the 4D SWS distribution of the Guinea savannah region of Ghana, which can support farmers in estimating where, when, how much, and for how long SWS is available for cultivation [1] . The data is useful for soil and agronomic research into crop yield production limited by water stress, such as modelling scenarios of water management for dry-season farming.…”

“…It provides useable information on the 4D SWS distribution of the Guinea savannah region of Ghana, which can support farmers in estimating where, when, how much, and for how long SWS is available for cultivation [1] .…”

Satellite-Based Estimation of Soil Moisture Content in Croplands: A Case Study in Golestan Province, North of Iran

“…For instance, a recent study demonstrated that the XGBoost algorithm is significantly more effective than random forest (RF) using Landsat 8 (L8) to estimate the SMC [22]. Machine learning techniques are increasingly widely used for predicting soil moisture using remote sensing data [23,24]. For instance, in a semiarid region of Iran, the SMC was estimated using machine learning algorithms [25].…”