1. Fish eat aquatic insects, and they can be sampled with benthic or drift sampling, above pollution and various points below. Various aquatic insect indices or species composition ratings can help evaluate changes above to below areas of pollution.

2. Most fish biologists I know use shocking of fish and netting to sample sections of streams or repeated sampling of habitats until no new species are found. Often the stunned fish can be measured and returned to stream with limited mortality.

3. Older, more destructive sampling technique was to net off sections of stream, add rotenone at upper section and capture all fish floating down to lower section, and add potassium permanganate to diffuse the rotenone toxicity. Mistakes sometimes have lead to extensive fish kills. This technique was sometimes used to remove non native, invasive species, and various permits and environmental assessment may be required if used.

4. Fishery biologists are trained to help design and quantify sampling data into population data.

5. As suggested, pollution types that produce known fish toxicity rates might be assessed with fluorescent dye studies as to dilution as water flows downstream. A fluorometer can be used to help quantify dilution as the dye flows downstream. Consulting a hydrologist would be helpful in designing dilution study or to help calculate dilution rates. Discharges of toxics during low flow periods are likely to have the least dilution with the greatest extent of impact. Conversely discharges of pollutants during high flows may have reduced extent due to the added dilution. Some pollutants are more mobile that others, as some may adhere, settle and bind with suspended particles or sediments or portions may be removed in collation or filtration in hyporheic substrate flow. Some pollutants breakdown quicker than others, so generalizing is difficult.

Some physico-chemical parameters such as temperature, pH, conductivity, BOD, DO, TDS and TSS can be use to evaluate impact of water pollution on fish population in a particular water body......

1. Fish eat aquatic insects, and they can be sampled with benthic or drift sampling, above pollution and various points below. Various aquatic insect indices or species composition ratings can help evaluate changes above to below areas of pollution.

2. Most fish biologists I know use shocking of fish and netting to sample sections of streams or repeated sampling of habitats until no new species are found. Often the stunned fish can be measured and returned to stream with limited mortality.

3. Older, more destructive sampling technique was to net off sections of stream, add rotenone at upper section and capture all fish floating down to lower section, and add potassium permanganate to diffuse the rotenone toxicity. Mistakes sometimes have lead to extensive fish kills. This technique was sometimes used to remove non native, invasive species, and various permits and environmental assessment may be required if used.

4. Fishery biologists are trained to help design and quantify sampling data into population data.

5. As suggested, pollution types that produce known fish toxicity rates might be assessed with fluorescent dye studies as to dilution as water flows downstream. A fluorometer can be used to help quantify dilution as the dye flows downstream. Consulting a hydrologist would be helpful in designing dilution study or to help calculate dilution rates. Discharges of toxics during low flow periods are likely to have the least dilution with the greatest extent of impact. Conversely discharges of pollutants during high flows may have reduced extent due to the added dilution. Some pollutants are more mobile that others, as some may adhere, settle and bind with suspended particles or sediments or portions may be removed in collation or filtration in hyporheic substrate flow. Some pollutants breakdown quicker than others, so generalizing is difficult.

https://www.sciencedirect.com/search?qs=fish%20population%20of%20a%20river&show=25&sortBy=relevance

Firstly, as answered above, there is bioindication analysis on microinvertibrates (they are not just a food for fish, but also a first indicators to pollution themselves)

Second, it is possibly to test the water in laboratory (acute or chronic test) on Daphnia Magna or Ceriodaphnia. Sensitivity of these biomarkers are close to sensitivity of fish spowning.

Also, it is necessary to do the phisico-chemical parameters of the water (pH, temperature, O2, turbidity)

Of course, it is possible to get and analyse the fish sample too, but it can be problematic. Bioindication and biotest usuall would provide a picture of the pollution influence to fish population.

Creel surveys are likely the oldest and easiest method of estimating the relative fish population size. Here you interview fishermen and look in their creel to see what they caught.

Catch per unit effort (fish per hour) can be estimated from the creel survey of fishermen by dividing total number caught by total hours fished. From this you can estimate trends in population.

There can also be historic sales figures for commercial fishermen. However, this is affected by the market price and honesty of reporting.

Mark and recapture is another classic method. https://en.wikipedia.org/wiki/Mark_and_recapture

Rotenone (fish poison) can be used if it's OK to kill some of the fish for counting.

There are probably many more methods. Details for using the methods I listed here can be found through a search.

https://www.google.com.pk/url?sa=t&source=web&rct=j&url=https://www.researchgate.net/profile/Peter_Bayley/publication/32963348_The_Fisheries_Analysis_System_FAS_Creel_Survey_Analyses/links/09e41506778d81502b000000/The-Fisheries-Analysis-System-FAS-Creel-Survey-Analyses.pdf&ved=2ahUKEwiFm5m5vM_hAhXC-KQKHT2cAIUQFjAKegQIBhAC&usg=AOvVaw1M9jkPAnSE4WGpLM9yngSx&cshid=1555241123968

https://www.google.com.pk/url?sa=t&source=web&rct=j&url=https://www.researchgate.net/publication/313949781_Creel_Surveys/amp&ved=2ahUKEwiFm5m5vM_hAhXC-KQKHT2cAIUQFjALegQIBhAI&usg=AOvVaw2wqmHOQLno4VR-Bf8juZJc&cf=1&cshid=1555241123968