In agriculture, land use and land classification address questions such as “where”, “why” and “when” a particular crop is grown within a particular agroecology. To date, there are several land suitability analysis (LSA) methods, but there is no consensus on the best method for crop suitability analysis. We conducted a scoping review to evaluate methodological strategies for LSA. Secondary to this, we assessed which of these would be suitable for neglected and underutilised crop species (NUS). The review classified LSA methods reported in articles as traditional (26.6%) and modern (63.4%). Modern approaches, including multi-criteria decision-making (MCDM) methods such as analytical hierarchy process (AHP) (14.9%) and fuzzy methods (12.9%); crop simulation models (9.9%) and machine learning related methods (25.7%) are gaining popularity over traditional methods. The MCDM methods, namely AHP and fuzzy, are commonly applied to LSA while crop models and machine learning related methods are gaining popularity. A total of 67 parameters from climatic, hydrology, soil, socio-economic and landscape properties are essential in LSA. Unavailability and the inclusion of categorical datasets from social sources is a challenge. Using big data and Internet of Things (IoT) improves the accuracy and reliability of LSA methods. The review expects to provide researchers and decision-makers with the most robust methods and standard parameters required in developing LSA for NUS. Qualitative and quantitative approaches must be integrated into unique hybrid land evaluation systems to improve LSA.

In this study, we conducted a scoping review and bibliometric analysis to evaluate the state-of-the-art regarding actual applications of unmanned aerial vehicle (UAV) technologies to guide precision agriculture (PA) practices within smallholder farms. UAVs have emerged as one of the most promising tools to monitor crops and guide PA practices to improve agricultural productivity and promote the sustainable and optimal use of critical resources. However, there is a need to understand how and for what purposes these technologies are being applied within smallholder farms. Using Biblioshiny and VOSviewer, 23 peer-reviewed articles from Scopus and Web of Science were analyzed to acquire a greater perspective on this emerging topical research focus area. The results of these investigations revealed that UAVs have largely been used for monitoring crop growth and development, guiding fertilizer management, and crop mapping but also have the potential to facilitate other PA practices. Several factors may moderate the potential of these technologies. However, due to continuous technological advancements and reductions in ownership and operational costs, there remains much cause for optimism regarding future applications of UAVs and associated technologies to inform policy, planning, and operational decision-making.

Developing and promoting neglected and underutilised crops (NUS) is essential to building resilience and strengthening food systems. However, a lack of robust, reliable, and scalable evidence impedes the mainstreaming of NUS into policies and strategies to improve food and nutrition security. Well-calibrated and validated crop models can be useful in closing the gap by generating evidence at several spatiotemporal scales needed to inform policy and practice. We, therefore, assessed progress, opportunities, and challenges for modelling NUS using a systematic review. While several models have been calibrated for a range of NUS, few models have been applied to evaluate the growth, yield, and resource use efficiencies of NUS. The low progress in modelling NUS is due, in part, to the vast diversity found within NUS that available models cannot adequately capture. A general lack of research compounds this focus on modelling NUS, which is made even more difficult by a deficiency of robust and accurate ecophysiological data needed to parameterise crop models. Furthermore, opportunities exist for advancing crop model databases and knowledge by tapping into big data and machine learning.

There is growing interest in promoting neglected and underutilized crop species to enhance agrobiodiversity and contribute to food systems transformation under climate change. A lack of available measured data has hindered the mainstreaming of these crops and limited the ability of agricultural databases to be used for calibrating and validating crop models. This study conducts a systematic scoping review and bibliometric analysis to assess the use of agricultural databases for crop modeling. The Biblioshiny App v4.1.2 and VOSviewer software v1.6.20 were used to analyze 51 peer-reviewed articles from Scopus and Web of Science. Key findings from this review were that agricultural databases have been used for estimating crop yield, assessing soil conditions, and fertilizer management and are invaluable for developing decision support tools. The main challenges include the need for high-quality datasets for developing agricultural databases and more expertise and financial resources to develop and apply crop and machine learning models. From the bibliometric dataset, only one study used modeled data to develop a crop database despite such data having a level of uncertainty. This presents an opportunity for future research to improve models to minimize their uncertainty level and provide reliable data for crop database development.

In light of a growing population and climate change compounding existing pressures on the agri-food system, there is a growing need to diversify agri-food systems and optimize the productivity and diversity of smallholder farming systems to enhance food and nutrition security under climate change. In this context, improving weed management takes on added significance, since weeds are among the primary factors contributing to crop yield losses for smallholder farmers. Adopting remote-sensing-based approaches to facilitate precision agricultural applications such as integrated weed management (IWM) has emerged as a potentially more effective alternative to conventional weed control approaches. However, given their unique socio-economic circumstances, there remains limited knowledge and understanding of how these technological advancements can be best utilized within smallholder farm settings. As such, this study used a systematic scoping review and attribute analysis to analyze 53 peer-reviewed articles from Scopus to gain further insight into remote-sensing-based IWM approaches and identify which are potentially best suited for smallholder farm applications. The findings of this review revealed that unmanned aerial vehicles (UAVs) are the most frequently utilized remote sensing platform for IWM applications and are also well suited for mapping and monitoring weeds within spatially heterogeneous areas such as smallholder farms. Despite the potential of these technologies for IWM, several obstacles to their operationalization within smallholder farm settings must be overcome, and careful consideration must be given on how best to maximize their potential before investing in these technologies.

Estimation of actual evapotranspiration (ET_a) based on reference evapotranspiration (ET_o) and the crop coefficient (K_c) remains one of the most widely used ET_a estimation approaches. However, its application in non-agricultural and natural environments has been limited, largely due to the lack of well-established K_c coefficients in these environments. Alternate K_c estimation approaches have thus been proposed in such instances, with techniques based on the use of leaf area index (LAI) estimates being quite popular. In this study, we utilised satellite-derived estimates of LAI acquired through the Google Earth Engine geospatial cloud computing platform and machine learning to quantify the water use of a commercial forest plantation situated within the eastern region of South Africa. Various machine learning-based models were trained and evaluated to predict K_c as a function of LAI, with the K_c estimates derived from the best-performing model then being used in conjunction with in situ measurements of ET_o to estimate ET_a. The ET estimates were then evaluated through comparisons against in situ measurements. An ensemble machine learning model showed the best performance, yielding RMSE and R² values of 0.05 and 0.68, respectively, when compared against measured K_c. Comparisons between estimated and measured ET_a yielded RMSE and R² values of 0.51 mm d⁻¹ and 0.90, respectively. These results were quite promising and further demonstrate the potential of geospatial cloud computing and machine learning-based approaches to provide a robust and efficient means of handling large volumes of data so that they can be optimally utilised to assist planning and management decisions.