No abstract

Vertisols are important agricultural soils in the Ethiopian highlands. The highland part of the Jama district is one of which Vertisols have huge coverage and are underutilized due to waterlogging. Such potential Vertisol areas need to be put under wise cultivation. Thus, a study was conducted to investigate the effects of soil drainage methods on surface runoff, soil loss, and yield of wheat crop as indicators of productivity improvement of typical Vertisol in the Jama district of Amhara Region, Ethiopia, during the rainy season of 2017/18. The treatment comprised three soil drainage methods (BBF120, BBF80, and BBF40) arranged in a randomized complete block design with three replications on standard runoff plots. Statistical Analysis System, version 9.0, was used to perform analysis of variance and mean separation of the collected data on yield, soil loss, and runoff. The result indicated that the effect of BBF120 brought significantly ( P < 0.05 ) higher difference on surface runoff, yield, and biomass of wheat over BBF40. The rainfall of about 55.05%, 51.45%, and 48.07% was lost as runoff from BBF120, BBF80, and BBF40, respectively. Drainage method BBF120 gave 34.6% and 17.3% of grain yield advantage over the drainage methods of BBF40 and BBF80, respectively, whereas soil loss was not significantly ( P > 0.05 ) different among all treatments; it is still in the range of soil loss tolerance in Ethiopia. As enhanced drainage is a requirement for successful crop production on Vertisol areas, BBF120 is recommended for draining excess runoff and consequently maximizing the yield of wheat in the study area and others with a similar farming system and agroecology.

Deep learning techniques help agronomists efficiently identify, analyze, and monitor tomato health. CNN locality constraint and existing small train sample adversely influenced disease recognition performance. To alleviate these challenges, we proposed a discriminative feature learning attention augmented residual (AAR) network. The AAR network contains a stacked pre-activated residual block that learns deep coarse level features with locality context whereas, the attention block captures salient feature sets while maintaining the global relationship in data points, attention features augment the learning of the residual block. We used conditional variational GAN (CVGAN) image reconstruction network and augmentation techniques to enlarge the training sample size and improve feature distribution.We conducted several experiments to demonstrate the AAR network performance.The AAR network performed 97.04% accuracy without data generation and augmentation, 98.91% with data generation and augmentation, and 99.03% trained with data augmentation, which consistently improved tomato disease recognition and visualization effectiveness in both cases by learning salient features than deep and wide CNN baseline networks and other related works. Therefore, the AAR network can be a good candidate for improved tomato disease detection and classification task.