Hello 

Can you provide length of river, possible source of contamination (Village discharge, industrial waste etc) and the mean depth of river then i can provide you information....

Regards

Siraj Goran

Without any idea on spatial and temporal variation, it will be hard to estimate it. How much knowledge do you have now on the river system?

Usually, one would take samples at a 'logical' spot, throughout the season, and determine the variation within and between the samples. Then you can work on the results and decide if you want to explain more or less variation.

The attached book will give you more details on designing a good monitoring scheme.

Chapter Sampling for Natural Resource Monitoring

Hi Vineet,

as Siraj has stated, we need a bit more information before we can give you a definite answer.

Of course it depends on the size and complexity of the river catchment - is it a single river course or are you monitoring a whole river network including tributaries?

But it also depends on the land use, the relief and the geology of your catchment. Is anthropgenic discharge (urban or industrial waste water, water from mining areas, agricultural drainage) the only source of pollutants or will there also be a geogenic component from the mountaineous parts of the catchment? Are there identifyable point sources (waste water treatment plants, drainage water dump sites)? Is there agricultural land use close to the river that could lead to a non-point pollution (nutrients) or to (bank) erosion (input of sediments which could e.g. contain heavy metals or phosphates).

Based on the characteristics of your research area you need to set-up your sampling network. Have one sampling point upstream of each point source and one sampling point immediately downstream of each point source. The same goes for the tributaries - have sampling points up- and downstream of each (major) tributary and a third one within the tributary (a kilometer or so upstream of the inflow point). In areas without point sources, have the sampling points spread out along the river and - if applicable - try to include different land use and relief types (if the river crosses a forest, have a point upstream and downstream of that forest and one or more points within).

Ideally you will end up with a dense network of sampling points that willk allow you to monitor the changes of the water chemistry, detect sources of pollution and e.g. evaluate the self-purification effects.

BUT, depending on the size/length of your catchment/river, this might not be feasible for you, as you might not have the logistical and financial capabilities to operate 20+ sampling points. Maybe you even only want to measure the summarized water quality at the outlet of a river catchment, so you only have a single sampling point?

Both is possible, as this is only the first part of the answer to your question - the spatial quantity. There is also, and at least as important, the temporal quantity of the samples:

How many samples do you need to take at each sampling point? Well, again, it depends. Do you have clearly defined point sources with a steady inflow of waste water (industrial or urban sewage)? Then maybe a couple of samples will be sufficient. But as soon as you have non point sources (agriculture) or highly variable climatic conditions (e.g. dry and wet season), you need a more systematic approach, as a single sample is just a snapshot of the water chemistry and you might just miss that crucial pollution peak by a day or two. How long do you plan the measure the water quality? For a year? Than I would go for at least 24 samples per point (two per month). Or just for a couple of weeks? Then go for daily samples (and make sure that the short duration you pick for your sampling is representative for the whole year).

The Bottom line is - the more samples the better, as more samples will help you catching those peaks without making them too prominent. Statistically speaking you need at least 11 samples before you can even start thinking about averages and such, but try to take as many samples as possible.

If necessary, find a balance between having more points and having more samples so that you do not get spread out too thinly in either of the two.

And keep in mind: Whenever you take samples from differnet points and want to compare them, collect some samples in as small a timeframe as possible in order to have comparable conditions (discharge, pollution, water temperature). This will lead to you driving around from point to point, quickly taking the samples, so that you have them all together within an hour or two. If you do not do that but rather take your samples for , e.g. May 2016 spread out from the 3rd of May (sample 1) to the 16th of May (sample 30) you will end up comparing apples and oranges as in between the 3rd and the 16th all kinds of things could have and will have happened (a heavy rainfall, some farmers applying fertilizers, a dramatic increase of the temperature etc.)

I hope that helps you getting started.

All the best,

Michael

Thank you all for your valuable suggestion. I am going to work on the River Ganga (Ganges). It is the third largest river in the world by the discharge. The Ganges Basin drains 1,000,000-square-kilometer (390,000 sq mi) and supports one of the world's highest densities of humans. The average depth of the river is 52 feet (17 m), and the maximum depth, 100 feet (33 m). Length of the river is 2525km. There are lots of sources of contamination throughout its course like big cities, industrial area, farmlands, STPs, Dams. The Ganges is the most sacred river of Hindus.  It is also a lifeline to millions of Indians who live along its course and depend on it for their daily needs. The Ganges was ranked as the fifth most polluted river in the world in 2007. More than 95% of the upper Gangetic Plain has been degraded or converted to agriculture or urban areas. The Ganga suffers from extreme pollution levels, which affect the 400 million people who live close to the river. Sewage from many cities along the river's course, industrial waste and religious offerings wrapped in non-degradable plastics add large amounts of pollutants to the river as it flows through densely populated areas

Hi Vineet,

with a catchment that large conducting field work yourself is sort of out of the question. You need to limit your sampling to a (small) part of the river (maybe 100 km long) and focus on the maybe most relevant/interesting section.

For the whole rest of the river, try to collect existing data (governmental water quality monitoring program, data from (inter)national research projects, ADB/UNDP/UNECE/Worldbank reports) and combine them into a metastudy. This task alone is probably worthy of a study and will be not easy to tackle, but it might provide you with a comprehensive analysis of the status quo. And based on those finding, you can select your research area for the actual fieldwork - if you still have the time and capacities for that.

Hi Michael , 

Actually, this is not my personal (single man) work, am working on a big project, so in this project we have to monitor the Physio-chemical parameter of a whole river including Heavy metal and Pesticides detection in the river, sediments and fishes. 

Allright, all the better.

Than you can stick to what I have suggested earlier, though organizing such a large monitoring network in such a way that all samples are taken simultaneously might be challenging. If you have enough funds, think about installing automatic sample collectors, that would make that part much easier.

But I would still advise making the collection and analysis of existing data a work package in your project.

I quickly checked how many stations are operated by the International Commission for the Protection of the Rhine:

There are 13 sampling points in the main river and 36 sampling points in the larger tributaries (see figure below). With a length of 1230km and a catchment of 185000km², the Rhine is approximately half the length of the Ganges and I guess that the number of tributaries is much larger in the Ganges catchment as well. Having around 30 main river sampling points and round about a hundred in the tributaries might be feasible. You will loose the granularity of measuring before and after each inflow, but with a catchment that large that is a acceptable compromise.

Dear Vineet Singh,

Statistically 30 samples is the best but depend on budget and time it could be reduce, but less than 10 do not recommended.

Dear Sayed Hadi,

Where do you base your 30 samples on? Space, number of tributaries, number of watersheds, seasonal variation?

@Vineet, it is a bit unclear what the purpose of your study is. Is this for a scientific valid sampling campaign, or for a project looking at some part of the Ganges and get 'an idea'?

I would strongly suggest, as Michael does, to start with searching for already published data, and then see what can be done. Without that, it sounds like any attempt to estimate number of samples should rather be based on budget than on scientific estimatons.

Dear Michel,

I said based on statistic that we need 30 sample for judging the results wit 95% confidence limit and if I said not less than 10 sample is in my experience.

Dear Seyed,

I would be interested to see your calculations, taking into account we are talking about the Ganges here, so we can discuss them in more detail.

You need to either use existing data or do a pilot study to determine both the within and between location variability.  You also need to understand the temporal variability.  If the variability is low then you need few sample if the variability is high you need more samples.  The most fundamental question is what differences matter to yo.  Is a change of 1% meaningful or is a change of 10% meaningful.  If your data are normally or log normally distributed then you can apply basic power analysis to determine your required sample size.  If not then the analysis becomes more challenging but can still be done.  Also,  you need to understand your detection limits and how you will deal with non detects which will occur.

Dear Michel Vorenhout,

The question is that how many sample is need to reach logic results. In this we could follow normal curve to get scientific result and in this we need 30 samples for the best result, but if there are lack of time and budget, then we could decrease the number to less that 30 samples.

You have many suggestions.  I think I can relate to Dr. Groll comments the most.  Basically, this could be a career project for a number of individuals, scientists, multiple millions of dollars, etc.  I know that is probably unrealistic.  But if this question deserves an answer, it needs to be partitioned into subject areas, and perhaps prioritized, as not all elements are in sync with each other, in time, space and frequency.  You also have a history (legacy) of pollution and probably hydrologic alterations such as dams, canals, diversions to consider.  Some chemicals such as pesticides may have temporary flushes while others may be more consistent, or show up when systems fail.  There are a lot of rabbits or side bars to follow to do a comprehensive job.  Several years back I bought a water quality sonde to address salinity, depth, temperature, and several other components in a wilderness area affected by man induced salinity.  But I wanted to take a boat trip up the lower Santee River in SC and define the salinity wedge, extent during a high tide (full or new moon) when we had salinity issues.  Consider what you might do with a portable water quality sonde on a cable, a suitable boat with depthfinder to navigate much of the river, hook into a GPS time recording system and start at the upper reaches of navigability and take a combination of sonde-time-depth_GPS synced cross sections, linear profiles on the navigable sections during a variety of conditions including perhaps a rain after the major pesticide treatments, during or after major floods when small dams that retain pollutants tend to fail, periods of low and moderate flow.  Since I don't know specifics about the Ganges River, I can only imagine from your dimensions of huge waters, with difficult sampling issues.  A pollutant in this large of a river is best measured as it enters or in some instances you much go downstream several to many meander bends so the chemicals are mixed across the  waters, but also diluted.  Anyone familar with fine sediments coming into a clear river has seen that that sediment hugs the tributary shore and slowly mixes.  Where you sample makes a difference as the sediment is  not mixed.  With a sonde on a long cable you should be able to sample the deeper navigable areas and a variety of basic water quality conditions, but likely not everything mentioned.  Whether sondes can be instrumented to cover all your areas of interest, is probably not possible as there may not be insitu recorders for some of these interests.  For those other interests, you will probably be taking some combination of water and sediment samples, and if your area has some quality biotic indicators and teams that can scuba dive, you may decide to use those such as macroinvertebrates as level of pollutants.  You need to map all the pollutant industries and land uses that you can, areas with faulty septic or poor sanitation issues, etc.  Of course you did not mention sideboards like budget, time, personnel, etc.  Realistically, you should reduce your question scope if there are too many side boards and a supportive team of experts.  I did not even get into the physio aspects or stream classification, morphology, sediment movement, channel stability, floodplain function, etc. but as I have said many times, Dave Rosgen's stream classification system and metrics (wildlandhydrology.com) are helpful to understanding and explaining to others stream and river functions, stability, adjustments, facets, etc.  As some have said, you need to know your purpose, expectations.  I can only imagine the adventure you may be on.  I think of John Wesley Powell (Colorado River explorer).

I suggest to you to read "Ecological Methodology" by Charles J. Krebs, 1999.

Dear Vineet,

Simple question with many different answers. For sampling stations follow Michael Groll, well explained plan. I suggest you to identify the sampling sites along the river. and decide how much area you want to cover or can cover. But I am worried. Is it an national program/ survey or study for PhD Degree. To make the study realistic many monitoring teams of scientists are required to start sampling at the same time on different pre-decided sampling sites along the river. You would required good monitoring team, field equipment and well established labs. if it is a national program you can train many young students during sampling, processing and analyzing. I do'nt know how long your program will proceed but at the end you will have huge data on physio-chemical parameter for the interpretation /discussion and conclusion. Many students can can also be enrolled for PhD degree program. you should also go through the previous studies if conducted earlier in your study or adjacent areas.

Good luck.

Hope it would help you to start

Javed Aftab, NIO Karachi 

I agree with Javed Aftab, moreover, what are the study aims and justification?

First of all, thank you all for your valuable suggestion. The project, I am just starting work on is a national level project; it’s a five year long project regarding the assessment of physio-chemical parameter of river Ganges including the role of heavy metals and pesticides and its effect on the fauna of River Ganges. There are no particular sampling sites; we need to cover whole Ganges. We have monitoring team, field equipment and well established labs. We basically focus on to check these parameters in different season  in every year. I know it’s a lot of work to do, that’s why I need your suggestion.