Smart Indigenous Weather App (SIWA) is an AI-powered climate intelligence platform that combines indigenous ecological knowledge with scientific weather data to provide localized, accessible 12–24-hour forecasts for smallholder farmers in Ghana. Trained on community-generated forecasting indicators and enriched with Ghana Meteorological Agency data, SIWA delivers actionable weather insights through an offline-first, local-language mobile experience. The platform strengthens farmer resilience to climate variability, reduces weather-related crop losses, and empowers communities, particularly women and youth, to become stewards of locally relevant climate intelligence. Through the ObaaSIWA project activities funded through the EmpowerHerAI project by the GRAIN project, over 100 female digital weather champions are being trained as part of efforts to provide timely weather information to female farmers who don't have access to smartphones
Background
In Ghana, 71% of agriculture solely depends on rainfall. This has been the norm for over many generations in many farming communities. As such, farmers have developed a natural intuition of forecasting the weather based on various ecological indicators such as star alignment, wind direction, temperature, plant flowering, seasonal behaviour of frogs/toads and some insects.
In late 2023, researchers from KNUST conducted a 10-month field study to scientifically evaluate the reliability of these indigenous forecasting methods in collaboration with 100 farmers across 14 Ghanaian communities in Assin Fosu (Central), Obuasi (Ashanti), and Kwabeng (Eastern).
Over the period and with over 100,000 forecast data collected from farmers, the study revealed that top forecasters reached 82% accuracy for daily forecasts, and 50% -86% correctly predicting major season rainfall onset. This result intrigued the researchers and the idea for Smart Indigenous Weather was born.
Problem Statement
Climate change and the prevalence of illegal mining (Galamsey) activities in Ghana have affected major farming communities across the southern regions. Galamsey destroys arable farmland, pollutes water bodies, and eradicates key ecological bio- and indigenous-indicators that smallholder farmers historically relied on to predict weather. At the same time, climate change has disrupted established rainfall patterns, causing unpredictable dry spells and false seasonal starts. Farmers plant according to ancestral calendars, only to be disappointed by failed rains and devastating crop losses, destabilizing already vulnerable livelihoods. Female farmers face additional barriers to accessing climate information, including limited access to resources, extension services and communication channels suited to their needs. This gender gap is particularly significant given women’s substantial contribution to agricultural production. SIWA addresses these challenges by making localized climate intelligence more accessible while positioning women as active participants in generating, using, and stewarding climate knowledge.
Objectives
SIWA was developed to achieve three main objectives:
Empower farmers with AI-driven climate intelligence they can rely on for their farming activities without disrupting their usual ways of carrying out these activities.
Empower women farmers – who contribute 52% of Ghana’s agricultural labour force, produce 70% of crops, yet own only 8% of agricultural land – as co-designers and stewards of climate data in their communities through the ObaaSIWA initiative funded under AI4D Africa, GRAIN project.
Enhance the economic self-reliance of marginalized smallholder farmers by reducing weather-related crop losses.
Description
Smart Indigenous Weather App (SIWA) is a climate intelligence platform that bridges the gap between traditional ecological knowledge and scientific meteorology.
It uses an AI model trained on indigenous ecological forecasting indicators documented from skilled local farmers over the course of the year. The community-derived forecasts are enriched with real-time scientific data from the Ghana Meteorological Agency (GMet) to produce localized 12 to 24 hour weather forecasts that significantly improve reliability and user trust.
The mobile app is designed to be highly accessible for low-literacy rural contexts through the use of integrated local language audio narrative, icon-driven interface and offline-first capability. Farmers make daily observations such as cloud patterns, heat levels, animal behaviours, etc. and receive actionable weather forecasts for their local communities.
Through SIWA, rural smallholder farmers are empowered to be stewards of climate intelligence in their communities. In pilot communities across Ghana’s Pra River Basin, app users have become community “weather ambassadors”, sharing predictions with neighbours and timing farm activities to reduce weather-related crop losses by ~30%.
Most importantly, SIWA extends beyond weather prediction. By digitizing traditional ecological indicators, the platform restores value to ancestral wisdom while driving community-led environmental advocacy. It raises local awareness of how climate change and destructive illegal mining (Galamsey) degrade vital bio-indicators and pollute ecosystems. Marginalized communities, particularly women and youth in agriculture, are empowered to take active ownership of their land, food security, and climate resilience.