Challenges

Achieving the zero hunger Sustainable Development Goal (SDG 2) is a huge challenge for most developing countries, especially those in sub-Saharan Africa. The Zambezi River Basin region of Africa has long faced severe food shortages due to low investments in agricultural technology and fertilizers, poor infrastructure and abnormal climate events, such as drought and flooding.

Timely and accurate agricultural information on crop growth conditions, production outlook and impact assessment of disasters is crucial for farm- ers and decisionmakers, yet it is generally difficult to obtain due to the lack of agricultural monitoring systems. To establish such agriculture monitoring systems usually requires large financial inputs, long-term maintenance costs, human resource commitments and advanced observation technologies, which constitutes a heavy burden for countries of the Zambezi River Basin. Cost-effective systems for crop monitoring and yield forecasting are needed in the Zambezi River Basin. In particular, cropland data at fine resolution is essential data for agricultural monitoring. However, such fine resolution data has been lacking in the Zambezi River Basin due to limitations of technologies and satellite data.

Toward a Solution

CropWatch, a remote sensing-based global agricultural monitoring platform integrates cloud computing and Earth Observation techniques, and can be used for monitoring and yield forecasting in developing countries. CropWatch was built by the Aerospace Information Research Institute (AIR) of the Chinese Academy of Sciences. The platform provides users with customized modules of agricultural monitoring information to carry out independent agricultural monitoring.

Under the facilitation of the United Nations Environment Programme (UNEP), the National Natural Science Foundation of China (NSFC), UNEP-IEMP, WorldBank and CropWatch4GEOGLAM, the project team developed a UNEP-NSFC joint research project for the Zambezi River Basin to apply CropWatch technology for cost-effective crop monitoring and yield forcasting. Since 2016, the CropWatch team, together with partners from Mozambique, Zambia and Zimbabwe, have carried out capacity building efforts in the Zambezi River Basin. Participants from Mozambique, Zambia and Zimbabwe attended a series of training courses in China and the Zambezi River Basin, during which they learned cropland mapping at fine resolution using machine learning meth-ods and high-resolution satellite data and how to use the CropWatch platform.

As the Zambezi River Basin spans a large area, a smart phone application and training course materials were distributed to local experts so they can collect field sample data in an efficient way. After the training, the CropWatch team worked with experts from the University of Zambia, the University of Zimbabwe, the Catholic University of Mozambique, the Ministry of Agriculture of Zambia and the Ministry of Agriculture and Rural Development of Mozambique to map the cropland in Mozambique, Zambia and Zimbabwe at 10-metre resolution using machine learning methods. In 2019, the dataset was handed over to the ministries of agriculture of the three countries and data support for agricultural cropping planning has continued. The data will help strengthen the ability of farmers and the countries to respond to abnormal weather and improve food security outcomes.

CropWatch was customized in particular for Mozambique on three aspects: language localization (Portuguese), adaptation of adminsitra- tive units and the monitoring period and model calibration according to local conditions. The customized CropWatch Cloud provides detailed information covering national, provincial, district and county administrative units. The models in the customized agricultural monitoring cloud platform were calibrated using locally obtained ground-measured data using the smart phone app to ensure suitability for Mozambique’s agricultural practices. Improved and adjusted CropWatch indicators, models and whole systems were adapted to Mozambique’s agricultural practices. The customized CropWatch Cloud and capacity building process not only improved the monitoring capacity of scientific and technical workers in Mozambique, but also promoted ownership and innovation of agricultural monitoring. It offers a unique opportunity for the country to monitor their agriculture without considering storage and computational resources.

This cloud platform solution can be replicated in other developing countries and regions as it provides customized agricultural monitoring systems, helping to improve agricultural monitoring capabilities and sustain agricultural development on a larger scale.