Soil information is vital for a range of purposes; however, soils vary greatly over short distances, making accurate soil data difficult to obtain. Soil surveys were first carried out in the 1920s, and the first national soil map was produced in 1940. Several regional studies were done in the 1960s, with the national Land Type Survey completed in 2002. Subsequently, the transfer of soil data to digital format has allowed a wide range of interpretations, but many data are still not freely available as they are held by a number of different bodies. The need for soil data is rapidly expanding to a range of fields, including health, food security, hydrological modelling and climate change. Fortunately, advances have been made in fields such as digital soil mapping, which enables the soil surveyors to address the need. The South African Soil Science fraternity will have to adapt to the changing environment in order to comply with the growing demands for data. At a recent Soil Information Workshop, soil scientists from government, academia and industry met to concentrate efforts in meeting the current and future soil data needs. The priorities identified included: interdisciplinary collaboration; expansion of the current national soil database with advanced data acquisition, manipulation, interpretation and countrywide dissemination facilities; and policy and human capital development in newly emerging soil science and environmental fields. It is hoped that soil information can play a critical role in the establishment of a national Natural Agricultural Information System.

This study aimed to identify the factors that influence smallholder farmers’ decisions to adopt four different conservation agriculture (CA) practices (i.e. minimum tillage, intercropping, cover cropping and crop rotation) in Mozambique. A non-probability sampling approach, incorporating both purposive and accidental sampling types, was followed. Three agro-ecological regions, followed by four provinces, were purposely selected. In addition, Accidental sampling was used to select 616 smallholder farmers from 38 communities in the three agro-ecological regions where CA projects were historically implemented by several NGO institutions. A questionnaire was administered to the 616 selected smallholder farmers. A descriptive logit model was applied in STATA to determine the probability of respondents adopting CA practices. The findings show that 44.6% of smallholder farmers adopted one or more of the CA practices, and 55.4% did not. It was also clear that most farmers did not adopt all components CA. Results obtained revealed that household size, animal ownership, communication assets (such as television, radio, and cell phone) and group membership had a positive influence on CA adoption. Interestingly, female-headed households were more likely to adopt CA. Awareness of soil health decline is an important factor determining adoption. The study concluded that the reasons for adoption are site-specific and a ‘blanket approach’ to promote adoption of CA is unlikely to be successful.

Heavy metals in water sources can threaten human life and the environment. The analysis time, need for chemical reagents, and sample amount per analysis assist in monitoring contaminants. Application of the Fourier Transform Infrared (FT-IR) Spectroscopy for the investigation of heavy metal elements has significantly developed due to its cost effectiveness and accuracy. Use of chemometric models such as Partial Least Square (PLS) and Principle Component Regression Analysis (PCA) relate the multiple spectral intensities from numerous calibration samples to the recognized analytes. This study focused on the FT-IR calibration and quantification of heavy metals (Ag, Cd, Cu, Pb and Zn) in surveyed water sources. FT-IR measurements were compared with the atomic absorption spectrometer (AAS) measurements. Quantitative analysis methods, PCA and PLS, were used in the FT-IR calibration. The spectral analyses were done using the Attenuated Total Reflectance (ATR-FTIR) technique on three river and four borehole water sources sampled within two seasons in QwaQwa, South Africa (SA). The PLS models had good R² values ranging from 0.95 to 1 and the PCA models ranged from 0.98 to 0.99. Significant differences were seen at 0.001 and 0.05 levels between the PLS and PCA models for detecting Cd and Pb in the water samples. The PCA models detected Ag concentrations more (˂0 mg L⁻¹ on selected sites). Both the PLS and PCA models had lower detection only for Zn ions mostly above 45 mg L⁻¹ deviating from the AAS measurements (<0.020 mg L⁻¹). The FT-IR spectroscopy demonstrated good potential for heavy metal determination purposes.