There are a few possibilities. Wildlife contributions - warm blooded animals, waterbirds, beaver, etc. Poor river mixing from source areas to sampling areas. Obtaining representative samples in river systems may require many samples across cross section and commensurate to velocity. Other sources unidentifiable. Possibility of ecoli multiplying — a maybe. It may help to also sample for fecal coliform, and compare results. Both are standard indicators of pollution, and they would not be considered good indicators if their numbers rapidly changed from one point to another. I believe I read a paper once asserting a possibility of ecoli reproducing outside of the animal gut under perhaps ideal conditions (temp, etc.), but don’t have anything to share.

Some rivers without turbulence will require a substantial distance downstream to achieve mixing, so pollutants can be delayed or hug the shore downstream. I coauthored a paper relative to fecal pollution in Stekoa Creek, a tributary of the Chattooga River. I moved a sample point downstream due to the mixing concern, based on my downstream observations of sediment discharges from Stekoa Creek, and how far downstream do you have to go before sediment plume was no longer discernible in the Chattooga River.

Storm runoff off can be a big contributor also.