Page 118 - EER-2-2

Explora: Environment
and Resource Climate suitability of AWD practice

practice where precipitation is <10 mm/day. However, the 5. Conclusion and recommendations
practice was appropriate in all dry and rainy seasons with
<20 mm/day of precipitation. Our findings correspond In this study, we have piloted climatic suitability analysis
with research from Burkina Faso, which reported that the of paddy rice cultivation and AWD technique for Eastern
Uganda, using the MaxEnt machine learning model
AWD technique is applicable during the drought and rainy combined with ENM in QGIS. Our research findings are
periods but also contributed to water saving (32% and 25%,
respectively) compared to traditional flooding irrigation. summarized below:
15
Considering the spatial variation of Pot , the spatial and (i) OCS, precipitation, accessible water (AW), TWI, the
pc
temporal precipitation distributions and ET significantly CRFVOL, and percolation rate collectively affect AWD
feasibility by influencing soil water retention, aeration,
affected the paddy water balance during the rainy season.
and nutrient cycling in paddy fields. These factors,
Conversely, studies in West Africa – Burkina Faso among others, greatly influenced the definition of
and Benin have demonstrated the potential of AWD for potential locations for AWD
improving water productivity. The research on a GIS-based (ii) AWD irrigation practice was climatically suitable in
climate suitability mapping in Burkina Faso revealed that paddy rice fields for all dry seasons in all evaluations
over 65% of the land was suitable for AWD during dry and locations with precipitation <10 mm/day,
15
and wet seasons. This aligns with our study (integrated percolation rates >5 mm/day, moderate OCS (10 tons/
machine learning and QGIS) findings, in which over 70% ha), and AW at a moderate level
of Eastern Uganda was suitable for AWD implementation, (iii) A sizable portion of the rainy season between March
provided precipitation levels remained below 20 mm/day. to June and September to November, corresponding
The study findings on climate suitability AWD irrigation to precipitation of 10 – 20 mm/day and 12 – 25 mm/
practice in Uganda have significant implications for day, was climatically favorable for AWD irrigation for
national irrigation policies to support paddy rice farming. paddy fields across the region
Integrating digital tools such as Internet of Things and AI (iv) AWD is unsuitable when the percolation rate is 1 –
with the AWD technique provides a robust sustainable 5 mm/day during the rainy season when precipitation
irrigation option to improve water management, rice is >20 mm/day. Understanding these factors enables
yields and mitigate greenhouse gas emissions. 8,49,50 This the optimization of AWD practice for precision water
underscores the need for policy interventions integrating management and to improve rice productivity
AWD into Uganda’s irrigation master plan for sustainable (v) The limitations of this study include the absence of
water allocation and wetland ecosystem management and accurate in situ data at groundwater level, percolation
services. 39 rates (calls for field estimation in the future), and lack
The global applicability of AWD stresses the potential of in situ validation.
for scaling up AWD technique in water-scarce regions. The study findings provide the first suitability assessment
However, implementation of this technique is hindered of AWD practice in Uganda and East Africa. MaxEnt and
by several challenges across regions: variations in soil ENM are suitable tools for estimating potential locations for
structure, water governance policies, technical know- paddy rice cultivation and AWD’s suitability with climate
how, and farmer adoption rates. The key technology change. However, future studies should consider using
component of AWD, mostly observation tubes, is ensemble modeling, which combines predictions from
affordable by smallholder farmers. Therefore, we multiple machine learning models (including random forest
7
suggest that the government direct its focus on training and artificial neural networks) to improve the accuracy
programs, field experiments, demonstration plots, and robustness of the suitability predictions. Predictions
50
and pilot projects, with extension services to help using ensemble technique like bagging (Bootstrap
farmers facilitate the adoption of AWD practice. AWD Aggregating) or boosting (AdaBoost, XGBoost) are more
applications require proper field planning and water accurate and stable. In addition, conducting country-wide
structures. Yet, the current irrigation infrastructures in comprehensive mapping and integrating groundwater
Uganda and East Africa are still underdeveloped, leading level data as a critical predictor supports AWD suitability,
to poor performance of most irrigation schemes. The expands spatial coverage, and ensures representation of
6
government should prioritize investments in irrigation the different agroecological zones and conditions. Our
modernization to enhance AWD implementation in findings are imperative for (i) assessing the suitability of
existing and new irrigation schemes to improve water use AWD practice and (ii) understanding the influence of AW,
efficiency and contribute to national food security while organic carbon, percolation rates, and precipitation on
promoting sustainable agriculture. optimizing AWD technique to improve irrigation planning

Volume 2 Issue 2 (2025) 14 doi: 10.36922/EER025040005