Stream dispersal of dead fish and survey effectiveness in a simulated fishkill

The dispersal and recovery effectiveness of tagged dead fish were determine d in a simulated stream fish kill. Two experiments were conducted under dif ferent flow regimes within a 2.3-km section of headwater stream. The first experiment used 120 tagged dead fish that were released at three locations when the mean daily flow was 153 L/s. In six surveys conducted over 2 d, we found that a majority of fish (58%) were dispersed less than 35 m. Fish th at were dispersed more than 35 m traveled quickly and traversed significant distances (up to 605 m). A unique distribution pattern was noted at each s ite, ranging from very limited to widespread dispersal for most fish. The s econd experiment used 216 tagged fish that were released at the same locati ons when the mean daily flow was 771 L/s. In six surveys conducted over 2 d , we found that 70% of the fish were dispersed more than 35 m. As in the fi rst experiment, a range of dispersal patterns was found. In both experiment s, surveys of tagged fish were effective, with 33 to 85% of fish found in e ach survey on the first day and with 81 to 93% of the released fish seen at some point during the experiments. However, daily surveys for dead fish ma y underestimate actual mortality, in part because there is substantial remo val by scavengers.