Understanding the dynamics and distribution of soil characteristics as influenced by landscape features and land use is critical for making decisions with regards to crop production and other types of proper land use. A field study was conducted to evaluate the relationship between topography position, land use and soil properties along the toposequence of Humbo Larena - Ofa Sere in Wolayita zone, Southern Ethiopia. Five slope classes were considered and a total of five pedons, one on each slope class, were opened and described at the study area. Soil samples collected from identified horizons of each pedon were analyzed for physicochemical properties. Additionally, random surface soil samples (0-20 cm) were collected from adjacent cultivated and grassland soils within a 100 m radius of each pedon; and three composites were made for each slope class. The pedons showed variability in physical, chemical and morphological characteristics of the soils. Field as well as laboratory determinations revealed the dominance of clay fraction in the soils. The existence of buried horizons with abrupt textural and sharp color changes showed the occurrence of lithological discontinuity in the pedon opened at depression. The available P content was higher in cultivated soils, whereas grassland soils had more organic carbon (OC), exchangeable bases and cation exchange capacity (CEC). Two soil types; Rhodic Alisols (Clayic) and Pellic Vertisols (Grumic) were identified in the area. Generally, slope and land use influenced soil properties, suggesting the need for different management practices for varying slope gradients and land uses for sustainable agricultural production. Key words: Toposequence, physicochemical properties, land use, pedon, slope gradient, landscape position, soil classification.

Maize (Zea mays L.) is an important food crop in Ethiopia, but productivity is low mainly due to low soil fertility and suboptimal fertilization. Therefore, this study aims to determine the yield, nutrient use efficiency and economic feasibility of maize production under various fertilizer applications and test the suitability of the Quantitative Evaluation of the Fertility in Tropical Soils (QUEFTS) model for predicting maize yield response to fertilization in Sidama region, southern Ethiopia. On-farm experiments were conducted at six sites (Site 1–6) of Sidama region, southern Ethiopia during the 2019 growing season. The experiments were laid out in a randomized complete block design (RCBD) with three replications. The experiment was a nutrient omission trial with seven treatments: control, two full NPK treatments and four nutrient omission treatments with contrasting N and P rates. Omitting N resulted in 5–28 % yield loss and omitting P resulted in 4–44 % yield loss compared to the lower rate of full NPK treatment across all study sites. Whereas omitting K resulted in 21 % yield loss only at sites 2 and 3. An increase in maize yield was mainly associated with an increase in both nutrient uptake and nutrient use efficiency of NPK. The results showed the need to revise blanket recommendations since the highest mean grain yields and net economic returns with acceptable marginal rates of return were obtained with NPK application or higher levels of NP (N2P2 treatment). In addition, the present QUEFTS model validation study revealed the good fit between QUEFTS model predicted maize grain yields (6.3 t ha⁻¹) to the average actual yields (7.4 t ha⁻¹) and this was also confirmed by small average values of RMSE = 1.5 t ha⁻¹ and PBIAS = 6.9 %. Thus, the model can be a promising option for development of site specific fertilizer recommendations under smallholder farming systems in the region.

In conventional furrow irrigated agriculture, excess water often causes leaching of nitrate-nitrogen. This experiment was aimed at quantifying nitrate leaching under farmers’ fertilizer and irrigation water use practices. It involved factorial combination of four rates of nitrogen fertilizer (0, 92, 184, and 368 kg N ha^-1) and two levels of irrigation (100% crop water requirement =1.00 CWR and farmers’ practice =1.25 CWR). It was conducted in onion (Allium cepa L.) planted drainage lysimeter at Melkassa Agricultural Research Center in 2015/16 cropping season. Nitrogen loss for the production season was determined from leachate collected every 10±3 days interval. Nitrogen uptake, dry matter (DM) accumulation, bulb yield, and bulb storability were measured. The result indicated low nitrate leaching for the season (< 22.46 kg NO₃^- ha^-1) however the losses were three-fold higher in high fertilizer rates or high irrigation water level. Independent t-test showed no significant difference (P ≥ 0.05) between irrigation levels on total N uptake, DM accumulation, and marketable bulb yield. Significant differences (P<0.05) were recorded between any different sub groups of N fertilizer rates on total N uptake except between 92 and 184 kg N ha^-1. The significant difference between 184 and 368 kg N ha^-1 on N uptake was not reflected in the DM accumulation and marketable onion bulb yield indicating low N use efficiency. The result showed about 48.5 to 69.4% surplus N accumulation in soil due to use of high rates of N fertilizer. Higher storage loss was depicted for application of 184 and 368 kg N ha^-1, and 1.25 CWR. Irrigating at 1.00 CWR and N fertilizer not exceeding about 184 kg N ha^-1 can reduce the problem without yield reduction. However, further field study is suggested within soils that have been under irrigated vegetable production for the last many years to reach at comprehensive conclusion and management recommendations.

Up-to-date digital soil resources information, and its comprehensive understanding, is crucial to support crop production and sustainable agricultural development. Generating such information through conventional approaches consumes time and resources, which is difficult for developing countries. In Ethiopia, the soil resource map that was in use is qualitative, dated (since 1984), and small-scale (1:2 M) which limits its practical applicability. Yet, a large legacy soil profile data accumulated over time and the emerging machine learning modelling approaches can help in generating a high-quality quantitative digital soil map that can provide accurate soil information. Thus, a group of researchers formed a coalition of the willing for soil and agronomy data sharing and collated about 20,000 soil profile data and stored them in a central database. The data were cleaned and harmonized using the latest soil profile data template and prepared 14,681 profile data for modelling. Random Forest was used to develop a continuous quantitative digital map of 18 WRB reference soil groups at 250 m resolution by integrating environmental variables-covariates representing major Ethiopian soil-forming factors. The validated map will have tremendous significance in soil management and other land-based development planning, given its improved spatial nature and quantitative digital representation.