Using data standards to build resilience in a warming world: lessons from environmental monitoring in Central Asia

By Eddie Crampton

Also available in Russian

Facing climate change and reducing its impact

The Amu Darya, Central Asia’s longest river, rises to its source in the mountains of Tajikistan. To the north, the Syr Darya flows westward from Kyrgyzstan to the Aral Sea. These rivers are fed by the seasonal melting of the mountains’ glaciers and together account for 90% of Central Asia’s river water.

Climate change is disrupting these seasonal patterns. Traditionally, glacier-melting begins in May and continues throughout the summer to provide a steady supply of fresh water in the dry season. Water from glaciers gradually feeds rivers in summer when rainfall is low, but when melting begins earlier in the year, river levels peak in spring and upset this pattern. This increases flood risk from glacial lakes and leaves less glacier ice to melt and provide for the hotter, drier summer when demand is high. Downstream, modern and Soviet-era canals redirect this river water through irrigation networks to support people across Uzbekistan, Turkmenistan, and Kazakhstan. As summer approaches and temperatures begin to rise, millions of people across the region will need dependable water supplies to flow down from the mountains in Tajikistan and Kyrgyzstan for drinking, sanitation, and agriculture. 

Efforts to mitigate the impact from climate change on these rivers are struggling because of a paucity of data. Monitoring agencies in these countries need visibility of conditions upstream to track glacier melting and spot disruptions to water supply in real time, information which is critical to coordinating decisions on water allocations, hydroelectricity, and responses to flooding. Water agencies need source-to-mouth data on glaciers and river levels. Monitoring programmes, however, aren’t coordinated across the region, so observations from researchers in the mountains cannot be interpreted quickly and used by those downstream. The fundamental problem here is a lack of interoperability between data sources: researchers measure ‘river conditions’ and track glacier melting in different ways; they store this data differently to one another; and they share their findings in formats that their counterparts cannot easily incorporate into their own work.

Lots of work is already being done in Central Asia to align policy and mobilise funding for joint-action in environmental monitoring. For this project to be successful, part of this joint action must be around data standards.

When rivers flow but data doesn’t: challenges of data fragmentation

In Central Asia, part of the challenge when it comes to coordinating collective climate action is regional data fragmentation.

Data fragmentation occurs when a group of people individually collect and store data on the same thing but follow different protocols and don’t measure what they’re looking for in the same way. For example, let’s imagine you were commissioned by your local library to study how many books were being borrowed by the public. Every week, you entered data into a spreadsheet: how many books were taken out, their genres, and how many pages each book had. After ten years, you decide to compare your local library to another. Fortunately, you discover I’ve been conducting the same study in my neighbourhood for the last ten years as well. I collected the same data but stored it all in handwritten notebooks. Another researcher, Paige, has been doing the same thing for her local library. Like you, Paige used a spreadsheet, but she noted down the weights of the books, how many words they had, and how long they were borrowed for. We have three useful datasets, but it would be difficult for us to share our findings with each other, partly because the data aren’t easily accessible and partly because we haven’t all collected and stored data in the same way.

In Central Asia, data fragmentation means organisations find it challenging to monitor river conditions outside of their immediate view. Just like in the example above, these nations do carry out excellent monitoring programmes, but the way they record and store their data isn’t aligned to a regional standard. This is what we mean when we say data isn’t interoperable – when you say you’re measuring ‘river conditions’ I need to know you’re measuring the same things as me and that your data is going to be stored in a way I can access. Sharing our findings on river conditions can only go so far when you’ve got spreadsheets full of observations on water levels, whilst I’ve got stacks of notebooks covering mineral deposits and water quality.

Data fragmentation isolates monitoring agencies and makes it harder to monitor glacier retreat, snowmelt, and fresh water supplies because each lacks a full view of river conditions elsewhere. Given they can’t use the data shared by other agencies, their understanding of real-time environmental risks is limited to the data they themselves are able to collect.

In the Soviet period, transboundary reliance on the Amu Darya and Syr Darya was managed through a centralised water-management administration, meaning upstream supply and downstream demand were jointly monitored. When the Soviet Union dissolved, this shared monitoring infrastructure dissolved as well. Authority was split across national agencies, marking the end of interoperability as shared practices ceased and datasets became siloed.

Reliance on these rivers hasn’t diminished, but the infrastructure that afforded a joint approach to monitoring conditions no longer exists. Now, long-term trends contain gaps and reporting formats aren’t designed for regional integration.

To counter this, organisations like Green Central Asia have developed models that simulate future water availability in basins like the Amu Darya. These models are a great resource for planners looking to anticipate long-term river conditions, but they don’t provide awareness of real-time risks.

Models like SWIM provide a view of what could happen to the Amu Darya in Uzbekistan if rainfall increased by, let’s say, 15%, but they can’t allow a researcher in Bukhara to see what is actually happening upstream right now. Instead, researchers use SWIM to visualise how changes in precipitation levels and glacier-melt timing will affect agriculture and hydroelectricity downstream, but models aren’t substitutes for a live-view of upstream conditions. In attempts to plan for prospective water availability, governments are increasingly having to rely on modelling to plug the gaps left by current monitoring that does not provide sufficient coverage of real-time conditions. With better visibility of glacier melt and water availability, governments could use models to simulate how real upstream conditions are likely to impact their communities.

When data is in sync, solutions become possible

So, where do data standards come into this? First, let’s look at what they actually are.

Data standards make our lives easier in a whole host of ways, though we’re rarely aware of how much we benefit from them. A data standard is essentially an agreement between people interested in working together for how a type of data should be stored and represented. They’re what allow your bank to understand the numbers on your payslip and are how my manager knew what I meant when I said I was posting this article on ‘29/04/2026’. Your bank and employer have agreed to read financial data in the same format, and my manager and I both understand what a date written as DDMMYYYY means.

Data standards are necessary because they enable us to work together in networks, from weather agencies sharing data with your phone, to global shipping, to making a dinner reservation via Google Maps. Without them, members of these networks have a really hard time understanding one another. It follows that the more people, organisations, or governments involved in a particular network, the more crucial data standards are to make collaboration possible.

In Central Asia, monitoring networks need to span borders, and the number of governments and NGOs engaged in these projects has the potential to be really high. Energy, agriculture, and livelihoods across five countries depend on coordination and regional alignment, and establishing regional data standards will bring these parties closer together.

Data standards would help to ensure interoperability in monitoring river conditions by aligning research practices at the outset and making the outcome of data-sharing and open-data practices more fruitful. With data standards in place, observations on snowmelt in Tajikistan could be applied to anticipating water allocations in Samarkand and understanding drought risk in Ashgabat. Standards would allow anyone monitoring river conditions to see the same picture, agree on conditions upstream, and understand risks the same way—no matter where they are. This makes coordination possible in responding to flooding, managing and anticipating water supplies, planning for hydroelectric projects, and building trust.

As mentioned, standards will also make models better. Where modellers and researchers are able to agree and verify that the underlying data are representative of what is actually happening upstream, models can be used more precisely by governments to plan for eventualities that they are more likely to confront. Looking forward, having better models also opens the door to incorporating AI into analysis and scenario-planning, enabling governments to improve and coordinate their responses to flooding and droughts. 

Central Asian nations are on the verge of a new era of cooperation, trust, and openness, in which collaborating on environmental monitoring can play a central role. These nations aren’t alone in this experience: two billion people rely on glaciers for water, with river networks crossing borders in South America, East Africa, and South Asia. The challenges that come with shared reliance on transboundary rivers are complex, and Central Asia has the potential to lead the way when it comes to the use of data in collaborative environmental monitoring.

Grid photo by Sergio Franklin on Unsplash