4.1. Fish Community Responses
Our results do not support our hypothesis that adding small instream wood at a density of 1 piece/m
2 would influence fishes in pools in channelized streams. Others [
30,
33] observed mixed fish community responses to the experimental addition of small instream wood bundles. Fourteen days after the addition of small wood bundles in the shallow sand-bottom littoral zone of a large floodplain river in Texas, fish abundance and species richness were reduced in locations with small wood bundles than reference locations without small wood bundles [
30]. These results were attributed to the lack of sunfishes (Centrarchidae), darters (Percidae), and catfishes (Ictaluridae) that are often associated with instream wood within the experimental reach and the lack of large instream wood near the experimental reach [
30]. Conversely, fish species richness and evenness were greater in reaches containing dense bundles of small instream wood than reference reaches lacking these small wood bundles in a channelized and incised third order sand-bed stream in Mississippi [
33]. Additionally, darters (Percidae) and madtom catfishes (
Noturus spp.) exhibited the strongest increases to the addition of the small wood bundles [
33]. These positive fish responses and localized changes in fish distribution were attributed to the complex cover provided by the small wood bundles and localized increases in the diversity of water depths and water velocities in the locations where the small wood bundles were installed [
33]. In light of these findings from the southeastern United States, we suspect that the lack of a fish community response we observed is a result of both the fish species composition within our experimental streams and the manner in which the small instream wood was added. Thus, 66% of all fish captures during our study were minnows (Cyprinidae), and only 33% of all fishes captured were sunfishes, darters, and catfishes. Thus, the dominance of minnows within our experimental pools may have hindered us from detecting an effect of adding small instream wood. Additionally, our small instream wood addition consisted of adding individual pieces of small wood placed haphazardly within each of the treatment pools. In contrast, both studies from the southeastern United States [
30,
33] used small wood and constructed instream wood bundles that were anchored to the stream bottom. Thus, these bundles, although constructed from small wood, are likely more accurately classified as instream wood accumulations and would be expected to influence the fish communities differently than separate individual pieces of small wood. Specifically, interstitial spaces provided by instream wood bundles and accumulations may provide better refugia from predators and have a greater impact on hydraulic conditions than small simple pieces of instream wood [
3].
In channelized streams in the Midwestern United States, fish communities have been observed to respond positively to instream wood additions [
2,
26,
27]. Angermeier and Karr [
2] conducted a split stream experiment and found that species richness, abundance, and the abundance of large fish were greater on the side containing instream wood compared to the side lacking instream wood in a channelized stream in Illinois. Angermeier and Karr [
2] also conducted a multiple reach experiment in the same stream, in which wood structures were added at different densities to treatment reaches and found that fish abundance was generally greater in the shallower treatment reaches than the deeper control reaches lacking instream wood. However, they noted that the effect of instream wood was often dependent on species and age class. Gatz [
26] increased overhead cover in a headwater stream in central Ohio by placing floating wood structures in treatment reaches that resulted in greater abundances of bluntnose minnow (
Pimephales notatus), creek chub, and longear sunfish (
Lepomis megalotis). Hrodey and Sutton [
27] added half log structures to three channelized streams in Indiana and observed greater abundance and species richness in reaches with half logs, but the frequency of positive fish responses to half log structures was dependent on the amount of pre-existing instream cover and adjacent riparian conditions. Positive fish responses to half log structures occurred more frequently in reaches lacking or having minimal amounts of instream cover and riparian zones dominated by row crops [
27].
Initially, we thought our results differed from the other Midwestern United States studies [
2,
26,
27] because the amount of instream wood we added was less than that added in other studies. The other studies did not use instream wood density as an indicator of the amount of instream wood added. Therefore, to facilitate comparisons, we calculated the percent of pool or reach area covered by instream wood in our experiment and those of Angermeier and Karr [
2], Gatz [
26], and Hrodey and Sutton [
27]. In our experiment, the treatment pools exhibited 3.9% instream wood coverage in treatment pools. In contrast, treatment reaches in Angermeier and Karr [
2] multiple reach experiment exhibited an instream wood coverage of 1% to 3%, treatment reaches in Hrodey and Sutton [
27] contained 1.7% instream wood coverage, and the treatment reaches in Gatz [
26] exhibited 3 to 4% instream wood coverage. Thus, we conclude that our results differ from other Midwestern United States studies because we added small instream wood pieces rather than large instream wood structures. Large and complex types of instream wood produce localized changes in flow that fish may use as refuge from stream current [
4]. Furthermore, large and more complex instream wood structures may support more individuals because they provide overhead cover for predator protection [
2,
26] and increased macroinvertebrate foraging area [
8].
In addition to providing a high-quality habitat, large and complex instream wood types remain effective longer because they are more resistant to environmental degradation. Of the 108 half log structures installed by Hrodey and Sutton [
27], 106 were still functional after two years. By the end of our field experiment, the added pieces of small instream wood had become covered with a thick layer of silt that may have reduced their effectiveness as a fish habitat. Although we did not observe a significant loss of added small instream wood pieces during our experiment, their small size and resulting high probability of downstream migration may reduce their ability to serve as fish habitats over longer time periods [
38].
4.2. Hydraulic Variable Responses
Our results support the hypothesis that adding small instream wood at a density of 1 piece/m
2 would not influence hydraulic conditions in pools within channelized streams. Our results are consistent with other studies [
2,
28,
29] that examined the influence of large instream wood on hydraulic conditions in channelized streams. During the split stream experiment of Angermeier and Karr [
2], water velocity slightly increased on the side lacking instream wood compared to the side with instream wood. Water depth and water velocity did vary between reaches with and without instream wood during the multiple reach experiment, but differences were not consistent and were attributed to the discharge regime rather than the addition of instream wood [
2]. Ehrman and Lamberti [
28] found that water velocity decreased in reaches that contained a channel spanning wood accumulation compared to reaches with wood accumulations only at the stream edge or reaches with no wood at all in an Indiana headwater stream. Additionally, Ehrman and Lamberti [
28] observed that discharge and water depth did not differ between reach types. Lester and Wright [
29] did not observe differences in water velocity and discharge between treatment and control reaches following an instream wood addition in agricultural streams in Australia. The degree to which instream wood influences hydraulic conditions is dependent on several factors, including size, structure, orientation relative to stream flow, and location within the water column [
29,
38,
39,
40,
41]. Our field-based hydraulic condition results combined with those of others [
2,
28,
29] suggest that instream wood types (i.e., small single pieces, large individual pieces of instream wood, large non-channel blocking wood accumulations, etc.) that do not create impoundments within agricultural streams may not significantly reduce the hydraulic capacity of these streams. Future research needs to document the effect of adding instream wood types that do not create impoundments on hydraulic capacity of agricultural headwater streams.