This type of sampling is often referred to as snapshot sampling. There should be plenty of examples in the literature. I remember that Chris Soulsby was amongst the first to use this type of sampling. I was involved in work in this direction too. You chould check out this paper:

Fröhlich, H.L., L. Breuer, H-G. Frede, J.A. Huisman and K.B. Vache. 2008. Water source characterization through spatiotemporal patterns of major, minor and trace element stream concentrations in a complex, mesoscale German catchment. Hydrological Processes, 22(12), 2028-2043.

But this is just one example.

Best,

Sander

Well,

You can apply the software like AQUICHEM, Surfer, GIS extension, Rockware, etc

Regards

Your example plot showing drainage area was impressive way to present. Since you have six drainages, you might try this for each parameter separate figures, but all six drainages combined into each parameter figure, keeping each drainage separate, no overlapping. Perhaps ArcMap might be helpful to accomplish this repeatedly for the varying parameters. But another approach would be use distance from headwater sample on x axis, and parameter value on y axis, color coded or separate symbols by the six drainage areas. It would be good to describe conditions of sampling as to baseflow, stormflow, etc., and someplace describe drainage sizes, geology/soils, stream type, land use within drainage areas, and any other conditions or differences meaningful in methods, etc.

In research at University level, how I was taught is we typically would consult with and receive advice or review from a statistician in design and planning stage preferable, especially if conducting a new approach. But no less important in this present stage of having data and how best to analyze and provide meaningful results would be discussing with an expert (statistician). Statistical analysis besides the parameter data may include some of the drainage or condition differences mentioned. Your idea of plotting the drainage, and color coding parameter change might be helpful in visualizing and describing differences associated with some flow regime sampled. Water quality typically changes with discharge rate, and as well can be different during rising stage of storm hydrograph and falling stage. So sampling during baseflow would be telling only part of the water quality story. In GIS, showing land use by drainage area may help explain some of these changes, and also good to show tributaries on plots even if not sampled as possible sources of parameter change.

Johan Alexander Huisman Many thanks! Yes I found quite a few studies on synoptic samplings in meso-scale catchments. What’s special in our dataset is the high resolution: up to 60 measurement points in catchments

Sorry I missed that! Looks interesting! Are their point sources or only diffuse sources? Upstream land use and area would be interesting explaining factors in case of diffuse sources. Good luck with the research. I always thought that synoptic sampling is very interesting and informative.

Best,

Sander

I forgot to add, barring flow data at each site, sometimes it can be useful to normalize data by drainage size. Each of your sampling locations has a drainage size associated with it. If perchance, you go downstream 50 feet, and one or more tributaries have entered the flow network, the amount of flow can substantially increase, as well as contributing area for nutrients, pollutants, etc. No normalizing by drainage size is dividing the parameter values by the associated drainage size. It can be accomplished with some work in Excel or ArcMap or other GIS software being used. When normalizing, the units can sometimes become strange, such as mg/l/square mile. If you had flow data instead of drainage size, you can take the concentration times the flow rate to get weight, such as 10 mg/l x 100 l/sec = 1 g/sec, as example. Each measurement site has a different drainage site and flow rate, so the flow factor or normalized drainage size factor would change. If the normalized units become unrecognizable, just describe how the data was adjusted.

Hi Remi, I did the same thing before. I have quite good longitudinal profile of DO and sac so on. For me, I used it to identify unknown pollution (point) sources and identify the DO decay rate in an urban river.

Hey Remi,

Sounds like a great project.

I'm pretty late, but I hope I can share a couple other references that could be of interest. Seems you've already looked into synoptic sampling (was going to suggest Robert Runkel's recent work with synoptic sampling in mine-drainage affected streams).

Something else interesting to check out could be the 'fast limnological automated measurements' project (http://lukeloken.weebly.com/flame.html). The project is more focused on lakes than streams but the ideas there could be helpful.

Looking forward to hearing about your results.

Dear Zach Simpson thank you for the hint, I didn't know this FLAME project!

Hi Remi, have a look at river network modelling, it may be the right tool for your data

Article Using spatial-stream-network models and long-term data to un...

Hi Remi, I'm working on a similar scale: 50 meter sampling over 1.5 km (DOC, EC) of a first-order stream in Northern Sweden. My work is sort of at the same stage now: trying to get a hold of emerging patterns. Curious to hear with which approach you end up!

A good reference for water quality transects and geostatistical models, focusing on kriging methods: Holbach, A., Wang, L., Chen, H. et al. Water mass interaction in the confluence zone of the Daning River and the Yangtze River—a driving force for algal growth in the Three Gorges Reservoir. Environ Sci Pollut Res 20, 7027–7037 (2013). https://doi.org/10.1007/s11356-012-1373-3