Amman — Algorithms capable of predicting air pollution before levels rise, identifying the likelihood of droughts and flash floods, forecasting temperatures before they occur, and simulating the outcomes of environmental decisions before they are implemented are opening the door to a fundamental shift in how Jordan manages its environmental and climate risks, moving from responding to damage after it occurs to anticipating it.
اضافة اعلان
However, this shift, which positions artificial intelligence (AI) as an early-warning mechanism rather than merely a tool for data analysis, raises questions about the Kingdom’s readiness to build a national system capable of predicting environmental problems. Challenges include data integration, the effectiveness of monitoring networks, institutional coordination, and translating scientific forecasts into actionable preventive measures.
Analysing massive amounts of data
AI is no longer simply a tool for analysing environmental data after a problem occurs. Globally, it is increasingly becoming a means of predicting problems before their effects become apparent, from forecasting air quality and floods to monitoring drought, changes in land use, and the potential effects of decisions before implementation, according to telecommunications and information technology expert Wasfi Al-Safadi.
Speaking to Al-Ghad, Al-Safadi noted that, under certain conditions, algorithms can analyse vast amounts of data, identify patterns that are difficult to detect through conventional methods, and generate probabilistic forecasts indicating what might happen, when and where.
This shift presents Jordan with an important opportunity, he said, while also raising broader questions about data, infrastructure, national capabilities, accountability, and the cost of building and maintaining such a system.
He stressed that AI does not “read the future” but calculates its probabilities. Modern models can learn from long time series containing weather, rainfall, air quality, water resources, satellite imagery, land-use data and other information, then identify patterns historically associated with particular risks.
As an example, he pointed to the European Centre for Medium-Range Weather Forecasts, which introduced its Artificial Intelligence Forecasting System (AIFS) into operational use in February 2025, alongside a conventional system based on physical models.
Trials demonstrated improvements across several forecasting metrics, as well as greater speed and lower energy consumption. However, AI does not replace experts or scientific models; rather, it adds greater capacity to analyse data, detect patterns and forecast probabilities, he explained.
In air-quality monitoring, the European Copernicus system uses machine-learning techniques to combine monitoring-station data with weather information and atmospheric models, improving forecasts for pollutants such as ozone, nitrogen dioxide and fine particulate matter, according to Al-Safadi.
Technically, he said, establishing a national environmental early-warning system is feasible, but it cannot be treated as a standalone AI application. It would require interconnected layers of data, including environmental monitoring stations; meteorological and hydrological information; air, water and soil quality measurements; industrial facilities and emission sources; satellite imagery and remote sensing; and land-use maps.
The AI layer would then integrate these sources and generate geographical and time-based maps of potential risks. Instead of merely indicating that an area is currently experiencing pollution, the system could show when the probability of pollution levels exceeding a specified threshold in that area has become high.
Jordan would not have to start from scratch, Al-Safadi emphasised. The national air-quality monitoring network currently includes 26 fixed stations across seven governorates. Meanwhile, Jordan’s National Artificial Intelligence Strategy for 2023–2027 includes 68 projects targeting sectors such as water, energy, agriculture, transport and smart cities.
The challenge, therefore, is not a complete absence of technology, but connecting existing data, institutions and capabilities into a unified system designed for prediction. AI cannot generate reliable forecasts from incomplete, irregular or incompatible data, he warned.
This creates a significant risk: a system may produce results that appear highly precise even though the actual level of certainty is much lower because of weaknesses in the underlying data.
AI can also evolve from an environmental warning tool into a decision-support mechanism. This is the principle behind simulation models and digital twins being developed worldwide to model Earth systems and test scenarios before decisions are made, Al-Safadi explained.
One example is the European Commission’s initiative to create a digital simulation of Earth, developed in cooperation with specialised European institutions. The initiative aims to build a highly accurate digital model using big data, high-performance computing and AI.
The initiative seeks to simulate natural phenomena, risks and the effects of human activity, while testing adaptation scenarios and policy options. Jordan could benefit from the underlying approach by developing a national-scale system tailored to the Kingdom’s needs, he said.
Al-Safadi stressed that if a system warns of a high probability of an environmental hazard and the relevant authority takes no action, the issue is not merely technical; it also concerns governance and legal accountability.
Conversely, if the system fails to predict an incident that subsequently occurs, the algorithm alone cannot be held responsible. The safest principle, he said, is to keep final decisions in human hands and ensure that forecasts can be explained, documented and audited.
This approach is consistent with modern international principles for AI governance, which emphasise transparency, explainability, risk management, accountability and human oversight.
Al-Safadi advised Jordan against attempting to build a massive system all at once. Instead, he recommended starting with a limited number of priority risks, including air quality, flash floods, drought and water scarcity, land degradation, heatwaves and changes in land use, before gradually expanding the system.
Estimating the cost of a national system
In Al-Safadi’s view, cost estimates should not be based solely on gross domestic product, population or the country’s geographical area.
He presented an analytical model based on eight criteria for estimating the scale of a local system compared with the European model. These include geographical area; population and settlements; economic size; topographical and hydrological complexity; cities and critical infrastructure; economic and environmental land use; climate complexity and risks; and the scope of the proposed system.
He noted that Jordan covers approximately 89,213 square kilometres and that its economy reached around $61.6 billion in 2025, according to World Bank data.
However, geographical area does not automatically determine cost. Sparsely populated areas may contain solar-energy projects, mining operations, roads, networks and strategic facilities, while more densely populated areas may require higher levels of monitoring and warning capability, he said.
Based on this multi-criteria analysis and using the European model as a technical reference rather than a budget that could simply be copied, Al-Safadi estimated that Jordan would need between $2 million and $4 million for an initial pilot phase focusing on a limited number of risks and making extensive use of existing infrastructure and data.
A national, multi-hazard system integrating data, monitoring stations and satellite information, developing models for air quality, water, drought and flash floods, and providing warning maps and decision-support tools for government agencies could cost between $5 million and $10 million.
An advanced system involving an expanded sensor network, multiple predictive models, scenario simulations, advanced computing infrastructure, a partial digital twin, and several years of operation, maintenance and upgrades could cost $10 million to $15 million or more.
Operating costs would be separate from the initial construction budget, Al-Safadi said. These would cover monitoring and sensors, data and computing, maintenance and model updates, cybersecurity, and staff training.
Turning early warnings into preventive action
Omar Al-Shoushan, president of the Union of Environmental Associations, believes Jordan has a genuine opportunity to move from managing environmental crises after they occur to anticipating and managing risks in advance.
This could be achieved by investing in an integrated national environmental early-warning system that uses AI, data analysis and scientific forecasts to protect people, natural resources and the economy, he said.
The issue is not simply whether the country possesses the technology, but whether it can turn early warnings into binding decisions and preventive interventions, Al-Shoushan stressed.
He called for a clear legal and institutional framework specifying who monitors a threat, who issues the warning, who decides whether to intervene, and what responsibility falls on an authority that receives a warning but fails to act. Early warnings should not remain mere reports or technical data, he argued.
Al-Shoushan believes Jordan has an important comparative advantage in this field because it has domestic companies and specialists with expertise in AI and digital technology.
This, he said, makes a locally developed early-warning system a feasible option and could create an opportunity for Jordan to become a regional centre of expertise in using smart technologies to strengthen climate and environmental resilience.
Investing in such a system should not be viewed as an additional burden on the state, but as preventive investment that could generate substantial future operational and economic savings. This is particularly important given limited resources, pressure on public finances and the international funding gap for adapting to the impacts of climate change, he added.
Early warnings allow governments to intervene when the cost of action is lower, rather than paying the price of damage and emergency responses after an event has occurred, Al-Shoushan said.
He urged that the system cover all parts of the Kingdom, prioritising water, agriculture, health, and productive and economic sectors to strengthen their resilience to environmental and climate risks.
Ultimately, he said, predictive information should become the basis for decision-making, shifting the approach from “repairing damage when it happens” to “anticipating risks and intervening before they turn into a crisis.” This, he concluded, is the essence of smart environmental and climate management.
Resource: Al-Ghad. Reporting: Farah Atiyat
