Search

• Biological control (BC) technologies are necessary to address the significant constraint that Fall armyworm (FAW) poses in cereal production in Africa. • The study employs a step-by-step modeling approach to map suitable sites of BC technologies, specifically parasitoids (Cotesia icipe). • Pest infestation levels are estimated using an evolutionary adaptive Neuro-Fuzzy inference system with a high coefficient of determination (R2) > 0.89. • We utilize fuzzy inference theory to accurately map the suitability of different regions in Kenya for the adoption of augmentative biocontrol using the parasitoids C. icipe in maize farms. • Artificial intelligence-based methods provide an effective advisory tool for guiding the deployment of biological control agents, such as parasitoids, for sustainable FAW management. Fall armyworm (FAW) Spodoptera frugiperda (J. E. Smith) is a major pest affecting cereal production in Africa. Biological control (BC) technologies are being promoted as a sustainable alternative to chemical control, which can lead to health risks and environmental hazards. However, effective deployment of these technologies requires site-specific recommendations. In this study, we use a step-by-step modelling approach to map suitable sites for BC technologies, focusing on the parasitoid Cotesia icipe using the FAW level of infestation, the ecological niche of the parasitoid, and the FAW host crop. The level of pest infestation was estimated using an evolutionary adaptive Neuro-Fuzzy inference system (R2 > 0.89) while the pest ecological niche was obtained using the maximum entropy algorithm (area under curve, AUC > 0.9). A fuzzy operator was used to combine all fuzzified variables into a single layer that represents the landscape's overall suitability for C. icipe in maize farms. Our computational findings indicate that C. icipe holds substantial promise as a BC agent in maize farms, with suitability levels consistently surpassing 90% throughout maize cropping seasons. The findings demonstrate that the utilization of artificial intelligence, combined with data science and knowledge representation, serves as an effective advisory tool for guiding the deployment of BC agents, such as parasitoids, for the sustainable management of FAW. This approach enables informed decision-making and enhances the efficacy of FAW management strategies by providing valuable insights and recommendations based on data-driven and computer intelligence analyses. [ABSTRACT FROM AUTHOR]

• This paper proposes and describes an interesting and potentially powerful means to validate large-scale biocontrol agent impacts during a Classical biological control (CBC) programme. • The methodology derives the beneficiaries' farmers from the dispersal of parasitoids at a landscape level. • The methodology was applied to the endoparasitoid Fopius arisanus release to control fruit fly Bactrocera dorsalis in Africa. • The study's outcome is appropriate for agriculture stakeholders and development partners to evaluate the benefits of CBC programme. Following the invasion of Africa by the oriental fruit fly, Bactrocera dorsalis , Classical biological control (CBC) has been exploited as a safer alternative for its suppression by the introduction and release of the koinobiont endoparasitoid, Fopius arisanus. Although the parasitoids have been released in several African countries, the extent of their dispersal and the resulting extent in benefits to fruit growers have not yet been elucidated. This paper proposes an innovative, multi-level CBC impact analysis that combines cellular automata (CA) and ecological niche models to estimate parasitoid dispersal ranges and the range of farmers who benefit from it. Firstly, we provided a generic, systematic methodological approach using CA rules incorporated into species distribution. Secondly, the model was used to estimate the dispersal range of the parasitoid, based on the life history and bioecology of the host insect (fruit fly) and the parasitoid. Finally, the parasitoid dispersal coverage was mapped across fruit crops attacked by the target fruit fly, and the number of households that benefited from the parasitoids release programme was extracted from the area of the dispersal. The data was projected across all countries where the parasitoid has been released and validated. In Kenya, the model showed that F. arisanus had covered a total area of 50.34 km2 from the initial point of open field release; and at the continental scale, the model predicted that the parasitoid had covered a total area of 229.97 km2. The model estimated that 351,855 and 3,731,330 households had directly benefited from the release of F. arisanus between 2013 and 2018 in Kenya and at the continental level, respectively. The study's outcome is appropriate for providing feedback information on the impact of CBC to government and development partners to assist in making informed decisions on technological interventions. [ABSTRACT FROM AUTHOR]

The performance of discrete mathematical models to describe the population dynamics of diamondback moth (DBM) (Plutella xylostella L.) and its parasitoid Diadegma semiclausum was investigated. The parameter values for several well-known models (Nicholson¿Bailey, Hassell and Varley, Beddington, Free and Lawton, May, Holling type 2, 3 and Getz and Mills functional responses) were estimated. The models were tested on 20 consecutive sets of time series data collected at 14 days interval for pest and parasitoid populations obtained from a highland cabbage growing area in eastern Kenya. Model parameters were estimated from minimized squared difference between the numerical solution of the model equations and the empirical data using Powell's method. Maximum calculated DBM growth rates varied between 0.02 and 0.07. The carrying capacity determined at 16.5 DBM/plant by the Beddington et al. model was within the range of field data. However, all the estimated parameter values relating to the parasitoid, including the instantaneous searching rate (0.07¿0.28), per capita searching efficiency (0.20¿0.27), search time (5.20¿5.33), handling time (0.77¿0.90), and parasitism aggregation index (0.33), were well outside the range encountered empirically. All models evaluated for DBM under Durbin¿Watson criteria, except the May model, were not autocorrelated with respect to residuals. In contrast, the criteria applied to the parasitoid residuals showed strong autocorrelations. Thus, these models failed to estimate parasitoid dynamics. We conclude that the interactions of the DBM with its parasitoid cannot be explained by any of the models tested. Two factors may be associated with this failure. First, the parasitoid in this integrated biological control system may not be playing a major role in regulating DBM population. Second, and perhaps more likely, poor correlations reflect gross inadequacies in the theoretical assumptions that underlie the existing models. [Copyright &y& Elsevier]

Abstract: The Lotka–Volterra model was applied to the population densities of diamondback moth (DBM), Plutella xylostella (L.) and its exotic larval parasitoid Diadegma semiclausum (Hellen) data that was collected earlier by icipe’s DBM biological control team. The collections were done for 15 months before the release and 36 months after release of the parasitoid in two areas; in Werugha, Coast Province of Kenya and Tharuni, Central Province of Kenya, respectively. For each area in pre- and post-release periods, we estimated Lotka–Volterra model parameters from the minimization of the loss function between the theoretical and experimental time-series datasets following the Nelder-Mead multidimensional method. The model estimated a reduction in the value of the steady state of DBM population from 4.86 to 2.17 in Werugha and from 6.11 to 3.76 and 3.45 (with and without exclusion of the time before D. semiclausum recovery) in Tharuni when transiting from the pre- and post-release periods, respectively. This change was a consequence of the newly introduced parasitoid, in the areas. The study presented a successful and detailed technique for non-linear model parameters restoration which was demonstrated by the correct mimicking of empirical datasets from the classical biological control with D. semiclausum, in different areas of Kenya. The applied model has measured the parasitoids impact on the DBM biological control through a quantitative estimate of the effectiveness of the newly introduced species D. semiclausum. These equations may therefore be used as tool for decision making in the implementation for such pests’ management system strategy. [Copyright &y& Elsevier]

Monitoring key pollinator taxa such as the genus Ceratina requires precise near real-time predictions to facilitate better surveillance. The potential habitat suitability of Ceratina moerenhouti was predicted in the Eastern Afromontane biodiversity hotspot (EABH) in Kenya using presence-only data, to identify their potential distribution and vulnerability due to climate change. Bioclimatic, edaphic, terrain, land surface temperature, and land use and land cover (LULC) variables were used as predictors. Three machine learning techniques, together with their ensemble model, were evaluated on their suitability to predict current and future (the shared socioeconomic pathways (SSPs), i.e., SSP245 and SSP585) habitat suitability. Predictors were subjected to variable selection using the variance inflation factor resulting in a few (n = 9) optimum variables. The area under the curve (AUC) and true skill statistic (TSS) were used for the accuracy assessment of the modeling outputs. The results indicated that 30% and 10% of the EABH in Murang'a and Taita Taveta counties are currently suitable for C. moerenhouti occurrence, respectively. However, future projections show a ±5% decrease in C. moerenhouti habitats in the two counties. Further, the ensemble model harnessed the algorithm differences while the random forest had the highest individual predictive power (AUC = 0.97; TSS = 0.96). Clay content, LULC, and the slope were the most relevant variables together with temperature and precipitation. Integrating multi-source data in predicting suitable habitats improves model prediction capacity. This study can be used to support the maintenance of flowering plant communities around agricultural areas to improve pollination services. • Monitoring key pollinator taxa facilitates better surveillance and the development of empirical interventions. • The potential habitat suitability of Ceratina moerenhouti was established in the Eastern Afromontane biodiversity hotspot. • Three robust machine learning algorithms and their ensemble had a very high prediction accuracy. • Integrating multi-source data improves the prediction accuracy. [ABSTRACT FROM AUTHOR]

• Fruit fly infestation data was modelled using Artificial Neural Network. • Fuzzy logic was used to transform plant physiology stages to numeric variable. • Fuzzy Neural Network can be used as a predictive tool for fruit fly population. • Prediction of insect pest infestation can allow for timely control of the pests. Avocado (Persea americana) production is increasing in Kenya, with both small and largeholder farming for domestic and export markets. However, one of main challenges that limit production is infestation by insect pests, notably the oriential fruit fly Bactocera dorsalis and Ceratitis spp. fruit flies, which cause direct crop losses and are indirectly responsible for non-tariff trade barriers due to stringent export requirements. Data on weekly pest trap counts were collected between September 2017 and December 2020 within orchards in avocado plantations. Fuzzy neural network (FNN) were used to model the population dynamics of B. dorsalis and Ceratitis spp. Weekly pest counts, rainfall, average temperature, relative humidity and avocado plant physiological stages were used for predictive modeling in different orchards. The performance of the resulting models was evaluated using coefficient of determination (R2), mean absolute error (MAE), mean relative approximation error (MRAE) and root mean squared error (RMSE). FNN models achieved satisfactory results in predicting the dynamics of the pests in the orchards, with most of the models obtaining R2 > 0.85. We demonstrated how FNN models can be used as predictive tools for managing and controlling fruit fly pest populations in these plantations, and how they may be suitable to predict fruit fly or other pests in similar cropping systems. Once the input variables are known, they can be loaded into the FNN models to predict field pest populations, and based on threshold values, allow for implementation of timely and adequate control measures such as the use of biopesticides. [ABSTRACT FROM AUTHOR]

In the Eastern Africa highlands, the gradual transformation of natural ecosystems to smallholding coffee-based agrosystems has resulted in more fragmented landscapes. Major pests of coffee find appropriate living conditions leading to high infestation rates and the need for smallholder farmers to implement pest control measures. This study aims to understand the influence of landscapes on the ecology of three major coffee pests: the coffee berry borer (CBB), Hypothenemus hampei , and the Antestia bugs Antestiopsis thunbergii (ABT) and A. facetoides (ABF). The study was conducted on a typical smallholder coffee-based landscape in central Kenya. The pest abundance was assessed monthly for two years in a network of 30 coffee plots spread across the coffee agro-ecological subzones (AEsZ), namely upper midland UM1 and UM2, and the transition zones between UM1 and UM2 and between UM2 and UM3, herein referred to as TZ1 and TZ2, respectively. Landscape metrics, viz. patch density, Euclidean nearest neighbour distance, proximity index, contagion index, interspersion and juxtaposition index were derived from a spatially explicit land cover map, based on 10 m Sentinel 2 data for nine buffer zones of radius ranging from 50 m to 1000 m around each sampled plot. Redundancy analysis (RDA) was used to establish the relationships between the observed pest abundances and landscape metrics, elevation, and AEsZ. Landscape indicators achieved the highest correlation with the pest abundances within a 300 m radius (Adjusted R2 > 0.5). Whereas beyond 300 m landscape scale, the predictor variables resulted in weak relationships (Adjusted R2 < 0.5) between the pests abundance and landscape metrics. We noted a strong influence of elevation and adjacency to cropland on Antestia bug populations. Specifically, ABF populations were negatively correlated with low elevation, whereas ABT's were positively correlated with high elevation zone. On the other hand, CBB was strongly influenced by contiguous coffee patches, especially in UM1 and UM2. Therefore, we recommend reducing connectivity between coffee patches for the management of CBB, whereas further studies should be conducted to identify secondary hosts of Antestia bugs that should not be adjacent or within coffee stands. • Landscape indicators achieved the highest correlation with pest abundances within a 300 m landscape scale • Antestia bugs were strongly influenced by elevation gradient and adjacency to cropland. • Coffee berry borer was strongly influenced by contiguous coffee patches within a limited landscape scale of 100 m • Reducing the connectivity of coffee patches will be useful in managing coffee berry borer. • Further studies should be conducted to identify secondary hosts of Antestia bugs. [ABSTRACT FROM AUTHOR]