In the first few months of 2022, the World Health Organisation (WHO) reported that the Horn of Africa region was experiencing its worst drought in decades. At the same time, weather anomalies worldwide were making international headlines. This year the UK shattered records with its hottest ever day, and in the first week of October villages across Pakistan were swept into monsoon-triggered floods.

The UN announced that worldwide this year’s July was one of the hottest ever recorded. It was among several organisations warning of the rising consequences of climate change, pinpointing key dates that have helped scientists to measure and compare impacts. This hard data drives home what scientists have been saying for years — that climate change is here to stay. The good news is that we can still prevent or mitigate severe damage in many areas.

Extreme weather, rising populations, and water scarcity have raised alarms about agriculture and food security across the Middle East and North Africa (MENA) region. But modern technology means that we can anticipate impacts days, weeks, and even months in advance. Satellite imagery and artificial intelligence are some of the tools making this happen.

Foreign corporations are increasingly seeing the potential for satellite-based and data-driven solutions in the MENA region in areas such as precision agriculture. “Between 2014 and 2020, a total of 33 investment deals in agri-tech start-ups in the wider MENA region attracted some $250 million in investment. However, a large portion of this was raised in 2020 in response to the pandemic,” noted a March 2022 report by the UK-based Oxford Business Group.

Non-profits have also taken an interest in aiding countries most affected by climate change. Al-Ahram Weekly spoke with Rachael McDonnell, deputy director-general at the International Water Management Institute (IWMI), a research organisation headquartered in Sri Lanka, about its work in the MENADrought project. For the past seven years, the team has been working with the governments of Morocco, Jordan, and Lebanon to develop drought action plans, notably using satellite imagery to combat climate change.

“The work started in 2015 in Cairo at a UN Food and Agriculture Organisation [FAO] meeting, where delegates from a number of different countries spoke about some of their most critical needs when it comes to managing water scarcity. Drought was identified by these three countries,” McDonnell said.

The US-funded project follows the three-pillar approach of the Integrated Drought Management Programme developed by the US National Drought Mitigation Centre at the University of Nebraska, one of the project’s main partners. Pillar one looks at early warning systems. Pillar two assesses impact and vulnerability. Pillar three focuses on management, mitigation, and response.

McDonnell noted that before this framework was implemented a drought would come along and an emergency meeting might be held. But by that point it would escalate into emergency responses and crisis management. Using the project’s “proactive management” method instead saves time, resources, and impacts because the key actors involved know what to do beforehand, she said.

“What we do know is that in the MENA region droughts are going to increase in frequency and intensity. We’ve seen this in Morocco, where there was a really bad drought. There was a little bit of recovery, but we’re now in the fourth year of moderate to extreme drought. It really is having major impacts now on agriculture and all sorts of water resources and the ability to provide water to different user groups in the country,” she said.

Drought is also not something that is defined by a single all-purpose number or metric. Rather, the definition is always relative. Average conditions vary according to locations but are measured by consistent indicators, such as rainfall and soil moisture. They are necessary to assess changes on region or country-wide levels, even narrowing measurements down to local areas.

“That’s why we use remote-sensing satellite imagery because what we can do is we can go back 20 years and see where we are at the moment in comparison. This is what it normally looks like in this particular month, or these are the normal conditions. Now it might be drier or wetter, and whatever is not normal we call it an anomaly,” McDonnell said.

AI: The project has also been utilising artificial intelligence (AI), more specifically a method to analyse visual imagery that is a type of Artificial Neural Network (ANN).

Artificial Neural Networks are computing systems commonly described as being modelled after the human brain. For instance, IBM defines it as process “mimicking the way that biological neurons signal to one another.”

At MENADrought, they have been using a specific class of ANN. “This is something known as ‘convolutional neural networks’ that can be used to analyse the existing global predictions. We can regionalise them and bring them down to the levels in a country. What’s fascinating is that in both Morocco and Jordan they are now being used by government officials because they have been shown to be pretty accurate. We’re now able to predict rainfall going up to three to four months ahead.”

Seasonal forecasting helps countries save money in the long run and to take these predictions into consideration when decision-making. Governments can decide whether or not to import wheat or ration food depending on factors such as projected crop and reservoir levels. McDonnell emphasises the necessity of accuracy whenever possible as human life and livelihoods are ultimately on the line. “When decisions are being made on that and you’re predicting the future, it’s quite a chance you’re asking people to trust that projection,” she said.

Many satellite and AI projects in developing countries rely on some degree of funding or partnerships from other countries. Even in projects unrelated to technology and climate change, foreign aid can sometimes be limiting, acting as a band-aid rather than a consistent treatment plan. But McDonell notes that from the start MENADrought sought to build self-sustainability in the countries they worked with.

“Drought technical committees have developed their own joint-action plans: they have installed services linked to the satellites they had there and have developed their own joint-action plans. We’ve facilitated and have had to ask questions, but they’re running the seasonal forecasting themselves and they’re using all of this already. We wanted to be able to walk away and have nobody even notice we’ve gone.”

 The flexibility of these systems and their relative ease of use make them more sustainable than many traditional methods. “These systems were designed to be operational, so we can put them into the Georgia Meteorological Department or the Morocco Ministry of Agriculture. We’ve designed these systems, so that they’re easy to use. The project is finished. But in all three countries, they are still continuing to generate the drought maps,” she added.

 

COP27: Islam Sediek is a computer engineer and entrepreneur who is also CEO of local start-up Meta Studio.

“Policy-makers and society at large rely on reliable forecasts to plan for the catastrophic impacts of climate change and to develop effective adaptation strategies. Artificial intelligence offers new ways that are potentially more accurate and faster than traditional weather and climate models for forecasting extreme events,” he said.

Meta Studio started two years ago and began focusing on climate change based on the urgency increasingly surrounding the subject. “A new method of deep learning has shown promising results for predicting complex systems. The power of deep learning is best achieved at a large scale, when models are trained on very large data sets. The process begins by collecting data about the weather then analysing it to predict future climate change and visualising it in a 3D model,” Sediek explained.

After learning more about the COP27 Conference, Meta Studio developed their own models to offer scientists more advanced technology to see accurate animation models rather than data projected onto 2D maps.

Egypt has its own set of challenges with data collection, which is why the models were developed with other countries in mind. Lack of historical data is one of the concerns Sediek has had, and rather than building from existing information, scientists and researchers often have to start more or less from scratch.

“Another point was the infrastructure of communication in several places. One device we required to collect data was drones. But this solution wasn’t available. Getting permission was also difficult for a start-up,” Sediek said.

Other challenges persist such as network and connectivity issues. And while governments and foreign-funded organisations often have the budgets to push through these limitations, start-ups and independent researchers may struggle to develop their own approaches. Sediek has faced a few of these issues, but he remains optimistic. The results from Meta’s existing models are also encouraging. “For now, our model can predict from a week and with 90 per cent accuracy, but if more data is added this will make the model predict with more time,” he said.

This is also where the AI comes in most notably in the prediction process. Once data is noted, the AI builds upon it in a process Sediek refers to as “training historical data”. This includes “temperature, humidity, pressure, and so on. While the AI draws these relations, it learns how to deal with this type of data.”

He hopes that policy-makers can use this data to lessen negative impacts and take action accordingly. Implementing satellite systems, AI, and similar technology is most successful when a variety of divisions are consulted and participating.

 

SATELLITE TECHNOLOGY: Juliette Hirsch is a project manager at Expertise France, also known as the French International Technical Cooperation Agency. This recently implemented a workshop in Egypt that sought to bring together diverse sectors in conversations about climate-focused satellite and AI tech. The “Greentech in Egypt D4D Workshop” focused on leveraging technology for environmental concerns.

“We chose two topics in relation to climate change: resilient agriculture and sustainable and smart cities. Not all parties are aware of the benefits of AI, and we want to do work with all the stakeholders involved — the public sector, but also the private sector, and not just companies, but also start-ups, incubators, accelerators, academia and civil society,” Hirsch said.

Policy-makers and stakeholders have already shown their interest in these subjects. Hirsch sees that different sectors and nations are on the same page and are creating mutually beneficial ties. “Egypt is already working on a lot of different partnerships that are very interesting and very promising. That’s something that we have been discussing together for a while, and so that’s why we came to those two topics together,” she said.

Satellite imagery is also an approach that Hirsch believes is gaining global notice. “We work a lot on spatial data, not just in Egypt, but also in other countries. And so that’s the topic of great interest for us,” she added.

Climate change is the type of subject that may not immediately appear as a “human” issue to the general public. But humanity is a key part of the discussion, from setting off and worsening climate-change triggers to battling these effects every day.

The value in developing better technology to combat climate change is especially needed because issues such as droughts do not exist in a vacuum. With countries recovering from the fallout of the Covid-19 pandemic, political conflicts, and economic challenges, the task can be especially daunting. Severe food insecurity, displacement, and unemployment commonly follow intense periods of drought. As is typically the case with crises, already vulnerable groups often suffer most. This commonly includes women, girls, and migrants that fall victim to human-trafficking when escaping drought-ravaged homes.

While discussing AI and agri-tech, Hirsch suggested always bringing the point back to those most affected by climate change and environmental challenges. It is something that the Expertise France organisers kept in mind before implementing the workshop. “Include associations that represent farmers or those for the safety of cyclists or people that are seeing impacts in their day-to-day lives and that are actually very implicated in these matters,” she said.

Another point she notes is that benefits can positively serve social and political incentives. By engaging different stakeholders, decision-makers can strengthen both the projects themselves and long-term ties between countries. “I think one of the workshop’s highlights was bringing together all the stakeholders around the table and developing a concrete roadmap of what can be done. Another highlight was to create potential partnerships between big companies and start-ups, along with the European ecosystem, Egyptian ecosystem, and African ecosystem,” she said.

Like Hirsch and Sediek, McDonnell also feels that the COP27 offers an exciting opportunity for conversations on AI and satellite-imagery approaches. The fact that these discussions have already started shows that the future for agri-tech is hopeful in Egypt and the Middle East.

As the MENADrought project packs up to move on to different countries, McDonnell reflects on progress happening in the MENA region. Egypt is now gearing up for the COP27, and it is not just the challenges that will take centre-stage.

 “This is a project with lots of lessons learned, but also successful projects in implementation. These can be taken to the COP27, where droughts and floods are going to be dominating the conversations. We have worked with really smart engineers and really clever policy-makers. We’ve had the support of ministers and secretary-generals across different ministries so we know that this can happen. Let’s build on that success,” she concluded.

---

*A version of this article appears in print in the 20 October, 2022 edition of Al-Ahram Weekly.