Central University of Technology’s ITIKI Project
A project incubated at Central University of Technology was successful in bridging recent technological innovation with indigenous knowledge, through the ITIKI computer science tool which can predict meteorological data inexpensively and accurately to assist local farmers.
Key Points:
- Droughts in Africa are increasing in frequency and severity, posing major threats to food security and livelihoods.
- Smallholder farmers often lack access to understandable, localized weather forecasts to help them mitigate drought impacts.
- ITIKI, a project incubated at Central University of Technology, is an innovative drought prediction system that combines indigenous environmental knowledge with scientific data and AI.
- By integrating traditional wisdom with technology, ITIKI delivers timely, accessible forecasts to farmers via mobile phones. The forecasts often come with actionable insights helping farmers make decisions based on all the data available.
In Zulu culture, they know that if the moon is “dirty”, it means a lot of rain is coming. This is true across the world, be it in Australia, Asia, Africa, Latin America. Small-scale farmers have used this indigenous knowledge. This knowledge is passed on from generation to generation and is not written down. My mother told me about it, and my grandmother told my mother about it. – Prof Muthoni Masinde, Central University of Technology.
Contents:
- 01. Introduction to ITIKI Project
- 02. By The Numbers: Africa’s Growing Drought Crisis
- 03. The ITIKI Approach: Bridging Traditional Wisdom and AI
- 04. ITIKI in Action: South Africa, Kenya and Mozambique
- 05. Traditional Knowledge: Example Weather Indicators
- 06. Project Impact and Future Applications
- 07. Frequently Asked Questions (FAQ)
Introduction to ITIKI
The growing frequency and intensity of droughts in Africa continue to threaten lives and livelihoods across the continent. The 2023 United Nations Global Assessment Report on Disaster Risk Reduction highlights an 80% increase in the risk of crop failure and hunger in sub-Saharan Africa.
Scientific models for predicting weather patterns and meteorological information are frequently communicated in ways that are not accessible or understandable to local farmers.
For example, forecasts might be presented in technical language or in a format that does not directly address the farmers’ immediate concerns.
Weather stations are also expensive to build and maintain, meaning many areas of rural Africa where crops are grown are not served with accurate localized data.
Enter ITIKI, an innovative project incubated at Central University of Technology which combined data and artificial intelligence with an underutilised resource; indigenous knowledge.
Prof Muthoni Masinde explained the inspiration behind the project saying, “I grew up in a typical Kenyan village and I remember going with my sister to fetch firewood. We heard a particular bird making noise, which in our local language, was interpreted as the bird saying “heaven come down.” When we heard that bird, we’d have to run home because it would rain precisely three hours later. It’s so precise.
“However, this indigenous knowledge is disappearing because young people have left the village and don’t want to be associated with it. They consider it primitive.”
Highlighting a primary challenge that ITIKI looked to tackle, Prof Masinde continued, “Climate change has also affected this knowledge. Farmers knew the seasons, like every March 28th it would rain, but it’s not the same anymore. They also knew that July would be cool, but now it doesn’t feel like winter anymore. So, this knowledge is not as reliable for the farmers.”
With climate change changing the reliability of indigenous knowledge but modern technology-driven indicators lacking localized data and input, could a combination of both approaches be more reliable?
By The Numbers: Africa's Growing Drought Crisis
- 3x. Droughts tripled: The frequency of dry and severely dry months in Africa has tripled since 1983.
- Landmass hit hard: In some regions of Africa, 40% more land is affected by drought compared to three decades ago.
- Millions at risk: An estimated 11 million children in East and Southern Africa face starvation due to drought-induced food insecurity.
- Crops devastated: Droughts can slash crop yields by up to 80%, significantly impacting food security and livelihoods across the continent.
ITIKI Approach: Bridging Artificial Intelligence and Traditional Wisdom
Understanding how ITIKI works requires delving into its innovative approach, which bridges the gap between traditional wisdom and modern scientific methods.
At its core, ITIKI is a drought forecasting system designed to assist smallholder farmers in Africa, who are often the most vulnerable to the impacts of drought.
The first step in ITIKI’s process involves collecting indigenous knowledge about environmental indicators. For instance, certain trees’ blooming patterns can signal upcoming weather conditions, and this knowledge is deeply rooted in the local communities.
These traditional indicators, passed down through generations, provide crucial insights into understanding and predicting weather patterns, especially in regions where scientific data may be scarce or unavailable.
Next, ITIKI enhances this indigenous knowledge with modern technology.
The system utilizes a network of sensors to collect data on various environmental factors such as soil moisture levels, temperature, and more.
This sensor data is a crucial component, offering real-time, objective measurements of current environmental conditions. The fusion of this data with traditional knowledge creates a robust base for forecasting.
ITIKI then processes this combined information through a complex event processing engine equipped with advanced AI algorithms. These algorithms are designed to analyze both sets of data, this step is critical as it transforms the raw data into actionable insights, tailored to the specific needs and locations of the farmers.
ITIKI delivers these forecasts directly to the farmers through messages sent to mobile phones.
In regions where large-scale, advanced farming technologies are not prevalent, and where mobile phone usage is widespread, receiving forecasts via SMS can be a game-changer.
ITIKI In Action
ITIKI’s solutions were implemented in three countries;
- Mozambique: Espugabera, Manica Province
- Kenya: Mbeere, Mbeere District
- South Africa: Pietermaritzburg, KwaZulu-Natal
ITIKI makes highly localized predictions by combining AI models with on-the-ground data. For example, the AI models can analyze decades of historical rainfall patterns in a specific village to forecast the upcoming rainy season. However, these models alone lack hyper-local precision. So ITIKI deploys low-cost sensors throughout the village to collect real-time temperature, soil moisture and other data. This bridges the gap, enabling predictions down to a 500 meter radius.
ITIKI also translates the forecasts into practical advice farmers can easily understand and act on.
As Professor Masinde explained to Techpoint Africa, instead of just predicting a rainfall amount in millimeters, ITIKI might say “The rainfall this season won’t be enough to grow your usual maize crop. Consider planting millet or sorghum instead.” This type of tailored guidance empowers farmers to adjust their practices to mitigate drought impacts.
By combining generational knowledge with cutting-edge AI, ITIKI achieves a blend of high-tech and time-tested insights. The result is actionable forecasts delivered in a language and format directly suited to each community’s needs.
AI still needs human expertise
The ITIKI project was launched in 2019. The adoption of Artificial Intelligence has accelerated rapidly since then yet in many applications of the technology similar challenges remain.
The successful on the ground implementation of ITIKI underscores a crucial aspect of technological advancement: the irreplaceable role of human knowledge, interaction, and communication.
In many contemporary applications, AI, despite its remarkable capabilities, falls short of providing genuinely practical input. This shortfall is particularly evident when AI systems attempt to operate in isolation from human context and expertise.
ITIKI’s success lies in integrating human expertise and needs into the technological process.
Incorporating indigenous knowledge also helps address some of the challenges that existing approaches to drought prediction possess.
Prof Masinde says “We should share scientific knowledge and indigenous knowledge to get the best of both worlds. That’s how my bridge was conceptualized. I use computer science tools, like affordable weather stations that are 10 times cheaper and can use phones as sensors. This brings the cost down significantly.
“Secondly, we’re using artificial intelligence, specifically artificial neural networks and fuzzy logic, to predict the weather.
“We can predict up to four years in advance with an accuracy of 70%, and up to 18 months with 98% accuracy.
“This is a huge solution because it means that millions of people who are at risk of hunger in Africa can be warned in advance.”
Traditional Knowledge: Weather Indicators
Researchers continue to gather information on which environmental indicators are used by local communities across Africa. Some examples of indigenous knowledge being applied to make farming decisions include;
| Location | Indicator | Meaning |
|---|---|---|
|
South Africa |
01. Birds building nests at higher locations 02. Early blooming of certain indigenous flowers |
01. Indication of heavy rains 02. Suggests an early start to the rainy season |
|
Tanzania |
01. Sudden increase in termite activity 02. Observations of the moon and stars’ positions and appearances |
01. Indicates impending changes in weather, such as the onset of rains 02. Insights into upcoming weather patterns |
|
Zimbabwe |
01. Changes in temperature and wind direction 02. Restless behaviour in domestic animals like cattle |
01. Indicators of approaching weather conditions 02. Hint at imminent weather changes |
|
Madagascar |
01. Appearance of rainbows, flowering of trees, behaviour of ants and frogs |
01. Forecast weather and climate conditions. |
|
South Africa |
01. “Dirty” moon (A hazy or halo-encircled moon) |
01. Significant rainfall expected. Farmers adjust planting schedules, stock up on water, and prepare drainage based on this observation |
Impact and Future Applications
ITIKI’s projects in Kenya, Mozambique and South Africa successfully empowered local farmers to combine technology with their local knowledge to make informed planting decisions.
Prof Masinde sees indigenous knowledge as a vital part of Africa’s farming future, saying” Our local knowledge, indigenous knowledge, is rich and authentic. It’s ours and no one can tell our story better than we can. I use this to tell my story, to reclaim and retell the story of African drought and hunger, as only we know it.
“We need to share our stories globally so that no one misrepresents us.”
Frequently Asked Questions (FAQ)
ITIKI uses indigenous knowledge related to environmental indicators, such as the blooming of specific trees, behavior of wildlife, and traditional weather patterns. This knowledge, passed down through generations, helps in predicting weather changes.
The primary beneficiaries of ITIKI are smallholder farmers in countries like Kenya, Mozambique, and South Africa.
These farmers use the forecasts provided by ITIKI to make informed decisions about cropping, thus improving their resilience to droughts.
Yes, ITIKI is designed to be accessible. Forecasts are delivered through SMS on mobile phones, making it suitable for farmers even in remote areas with limited access to advanced technology.
While ITIKI is currently focused on African countries, the underlying concept of integrating indigenous knowledge with scientific methods can potentially be adapted to other regions, subject to the availability and relevance of local traditional wisdom.
The future of ITIKI involves expanding its reach to more African countries, refining its prediction models, and possibly adapting the system for other environmental challenges beyond drought forecasting. The project aims to continue enhancing the resilience of smallholder farmers to climate-related challenges.
