PhD Pilot Blog

What If Water Monitoring Didn’t Take Weeks? Bringing Field Measurements straight to the Map

Mikhail Sandzhiev, University of Eastern Finland, misandzh@uef.fi


Water quality monitoring is essential for protecting ecosystems and human health, though the path from sampling to actionable information is often slow. In this blog, I explore how portable field measurements, automated data processing, and GIS mapping can be combined into a single digital workflow- transforming raw metal concentration data into interactive maps within minutes instead of weeks.

Field-deployed p-XRF system for on-site metal monitoring in a Finnish lake environment during the site validation.
Photo 1. Field-deployed p-XRF system for on-site metal monitoring in a Finnish lake environment during the site validation. Photo by Mikhail Sandzhiev. 

From Numbers to Maps

What if water quality data did not sit in a laboratory report for weeks before anyone could use it? What if metal concentrations measured in the field could appear instantly on a map, ready for interpretation?

Traditional monitoring of metals in surface waters relies on laboratory techniques such as ICP-MS. These methods are precise and reliable, but they require sample transport, preparation, and specialized analysis. The result is often a delay between sampling and usable information. In dynamic aquatic environments, where conditions can vary spatially and temporally, this time gap limits our ability to react quickly. In my doctoral research within the Digital Waters (DIWA) Flagship, I explore how we can shorten that gap by digitalizing the entire workflow – from field measurement to spatial visualization. Instead of separating sampling, analysis, processing, and communication into disconnected steps, we connect them into one continuous system.

The core of the approach is portable X-ray fluorescence (p-XRF) combined with a metal concentration filter. A water sample is pumped through a small filter that collects dissolved metals. The filter is then analyzed directly on site using a handheld p-XRF instrument. Within minutes, we obtain quantitative information about several metals simultaneously.

Figure 1. Custom 3D-printed holders (a) for preconcentration of the BP-TCPSi filters
Figure 2. (b) for mounting the BP-TCPSi filters for p-XRF analysis. 

However, measurement alone is not enough. Numbers in a spreadsheet do not automatically support environmental decisions. This is where digital processing becomes essential. A custom Python script converts the raw instrument output into structured, georeferenced datasets. The concentrations are automatically processed, validated, and formatted into GIS-compatible files. These files are then imported into QGIS, where each measurement is linked to its geographic location.

To simplify interpretation, we also apply a Water Quality Index (WQI). Instead of examining five separate metal concentrations one by one, the WQI summarizes the overall chemical status into a single value. Importantly, the individual metal concentrations remain accessible in the attribute table for detailed inspection. The map provides an overview; the data remains transparent underneath.

The result is a dynamic map where lakes are symbolized according to their water quality status. By clicking on a lake, the user can instantly view the measured concentrations, the calculated index, and relevant spatial metadata. What previously required laboratory reporting and manual data handling can now be achieved in a fraction of the time.

27.2.2026

Share the Post:

Related Posts

PhD Pilot Blog Iiro Seppä

Spätzle, Strudel & Neural hydrology: My research visit to Linz

Iiro Seppä, University of Turku. iielse@utu.fi At the end of November, I spent two weeks with the Machine learning in Earth Sciences group at Interdisciplinary Transformation University (IT-U) in Linz, Austria. Photo 1. Researcher Iiro Seppä enjoying the weekend in Hohensalzburg, with Gaisberg in the background. Photo by Jasmin Toukoniemi.

PhD PIlot Blog Raffa Ahmed

Studying Hydropower in Cold Climates: What I Learned About Resilience

Raffa Ahmed, University of Oulu, Raffa.Ahmed@oulu.fi When I began my PhD, I thought resilience was a technical concept, something engineers talk about when designing turbines. But as I dive deeper into hydropower in cold climate regions, I have learned that resilience is about much more. It is about systems, people,