Mechanisms Limiting a Vertebrate Invasion: Brook Trout in Mountain Streams

I used multiple approaches to study brook trout invasions in mountainous streams of Idaho and Montana, USA. After studying marked fish in experimental reaches and fish distributions throughout a drainage, I concluded that larger brook trout (95mm) moved more than smaller ones and were probably responsible for most dispersal leading to invasion. Contrary to earlier hypotheses, poor swimming ability did not prevent brook trout from ascending streams with channel slopes of 13 % over 67 m. During the summer, brook trout moved upstream more than downstream even in the steepest experimental areas. Nearly vertical drops in steep reaches rather than steep slopes per se, prevented upstream movements. In downstream-directed invasions (originating from headwater lakes), brook trout apparently dispersed downstream through 80 % slopes and over 18-m-high waterfalls and occupied steeper stream channels than they appear capable of ascending. In the South Fork Salmon River drainage, Idaho, little change in upstream brook trout distribution limits occurred over 25 years. In the three streams where invasion occurred, the average apparent invasion rate was 5 to 11 m per year, suggesting that invasion may have occurred in pulses. Dispersal ability clearly limited invasion in several streams, but not in others. Invasion of accessible habitat is not necessarily inevitable in the short term (several decades), but because brook trout can ascend steep streams, more area is accessible than commonly assumed and may ultimately be invaded. Brook trout annual growth rate (measured by length at age) declined upstream in a non-lake-fed stream but increased upstream in a lake-fed stream and was correlated with stream temperature in age classes 0 and 1. Later maturation and reduced fecundity associated with slow growth were sufficient to substantially reduce upstream subpopulation growth rate unless compensated by increased survival. Life-stage simulation analyses indicated that early age class survival rates and early maturation had strong effects on population growth. In the non-lake-fed stream, patchier recruitment of age 0 fish and slower growth upstream likely contributed to creating the distribution limit. Patterns and processes observed were consistent with the idea that source-sink dynamics help to sustain distributions of brook trout in streams.