Lateral Movement of Ground‐Dwelling Invertebrates Resolves Discrepancies Between Fish Diet and Terrestrial Input Data

ABSTRACT Terrestrial invertebrates are important food resources for fish in forested streams, and understanding their origins as fish prey—whether from air‐canopy or ground surface‐soil environments within the forest—is essential for the management of headwater catchments. However, the contribution of ground‐dwelling invertebrates to fish food resources remains unclear because few studies have quantified their inputs to streams. Here, we investigated the input pathways and composition of terrestrial invertebrates supplied to a channel and preyed upon by fish in a small forested stream in a cool‐temperature area of western Japan and evaluated the importance of ground‐dwelling invertebrates as a food resource for stream fish. We used pan‐traps placed in the central part of the channel (instream traps) and along the channel margins (bankside traps) to clarify the input pathways of terrestrial invertebrates (vertical pathway via air and canopy; lateral pathway via ground surface and soil). Additionally, terrestrial invertebrates were identified at the class, order, and family levels to determine their typical habitat and compare mass‐based composition in terrestrial inputs and fish stomach contents. The mass of terrestrial invertebrates captured by traps at the channel margin was equivalent to approximately 10 times that in the central part of the channel, mainly due to ground‐dwelling invertebrates common in bankside traps but rare in instream traps. Based on an assessment at the reach scale, the proportion of ground‐dwelling invertebrates to total terrestrial inputs (58%) was comparable to their proportion in fish stomach contents (58%). These findings indicate that bankside traps effectively address the methodological limitations of relying solely on instream traps and highlight ground‐dwelling invertebrates as an essential component of potential fish prey. However, the correlation between the inputs and stomach contents was notably weaker for ground‐dwelling invertebrates ( r = 0.47) than for air‐ and canopy‐dwelling invertebrates ( r = 0.78). This weak correlation might stem from the complex prey–fish processes, such as the spatial heterogeneity of ground‐dwelling invertebrate inputs, their underwater behaviours, and/or fish prey selections. Our results emphasise the need for quantitative analysis of the lateral movement of ground‐dwelling invertebrates to better understand forest–stream ecological linkages through terrestrial invertebrate inputs. Therefore, we conclude that conserving the forest floor and soil environment in riparian areas, as well as vegetation, is crucial for maintaining stream ecosystems.