Effects of varying tillage practices and weed control methods on the efficacy of infiltration models.

Agricultural land preparation and weed control techniques are essential farm management tools that affect the dynamics of soil water infiltration and the estimation accuracy of infiltration models. To analyse the interaction effect of tillage and weed control methods on the changes in soil physical properties and the efficacy of infiltration models, an experiment was conducted on a sandy clay loam forest ochrosol at Hodzo near Ho in Ghana. Four tillage systems (No Tillage [NT], Reduced Tillage [RT], Plough + Harrow + Ridging [PHR], and Deep Tillage + Plough + Harrow + Ridging [DPHR]) and three weed control methods (Hoeing [H], Machete [MAT] and No Weeding [NW]) were employed. The study also tested the reliability of the models (Kostiakov, Philip, and Horton) using the goodness of fit statistical criteria: Root mean squared error (RMSE), Mean absolute error (MAE), Coefficient of determination (R2), and Nash-Sutcliffe efficiency (NSE). The results show that conservation tillage systems (CsT) and conventional tillage systems (CT) with MAT weeding treatments recorded the highest moisture content across the studied soil profile, especially for NT x MAT (11.189%) which was significant (p < 0.05) in the 15-30 cm layer; the lowest were observed in the CsT and CT with H weeding interactions, especially for the DPHR x H (8.086%). Comparing the interaction effect on the soil infiltration, the highest mean infiltration rate was significant (p < 0.05) under the NT X H treatment combination whilst the lowest infiltration rate was recorded in the DPHR X H and PHR X NW treatment combinations. The efficiency of the fitting models (Kostiakov > Horton > Philip) highly prioritised the soil tillage operations and weed management under the treatments DPHR x MAT > DPHR x NW > DPHR x H > RT x MAT > PHR x NW > PHR x MAT > NT x NW > RT x MAT > PHR x H > RT x H > NT x MAT > RT x NW > NT x H in that order. The trend shows that the increase in tillage intensity and the decrease in weed management intensity induce the quality of the estimation process and vice versa. The study, therefore, identified the use of machete (MAT) with DPHR under the Kostiakov model as the efficient land management for modelling the cumulative infiltration characteristics of the sandy clay loam ochrosols of the study area.

Land Cover Changes Utilising Landsat Satellite Imageries for the Kumasi Metropolis and Its Adjoining Municipalities in Ghana (1986-2022).

Forest loss, unbridled urbanisation, and the loss of arable lands have become contentious issues for the sustainable management of land. Landsat satellite images for 1986, 2003, 2013, and 2022, covering the Kumasi Metropolitan Assembly and its adjoining municipalities, were used to analyse the Land Use Land Cover (LULC) changes. The machine learning algorithm, Support Vector Machine (SVM), was used for the satellite image classification that led to the generation of the LULC maps. The Normalised Difference Vegetation Index (NDVI) and Normalised Difference Built-up Index (NDBI) were analysed to assess the correlations between the indices. The image overlays of the forest and urban extents and the calculation of the annual deforestation rates were evaluated. The study revealed decreasing trends in forestlands, increased urban/built-up areas (similar to the image overlays), and a decline in agricultural lands. However, there was a negative relationship between the NDVI and NDBI. The results corroborate the pressing need for the assessment of LULC utilising satellite sensors. This paper contributes to the existing outlines for evolving land design for the promotion of sustainable land use.