<p>Disturbance regimes like freezing and thawing (FT) can have potentially significant impacts on nutrient release from soil and are predicted to increase with climate change. This is particularly important in biogeochemical hotspots like floodplains that can both remove and release nutrients to surface waters during flooding. Connection between the river and floodplain can improve water quality by reducing nutrient loads through microbial processes and sedimentation. However, conditions during flooding can also lead to phosphorus (P) release from pools that are not normally bioavailable. Disturbance events like FT can also lead to changes in bioavailable P due to microbial cell lysis. This study investigates differences in P chemistry and flux during flooding from intact soil cores that have undergone a FT cycle compared to soils that have not undergone freezing. Floodplain soils were collected from four sites along the Wabash and Tippecanoe Rivers in Indiana. We hypothesized that (i) the primary pools of P within the soil would change with freezing (ii) and flooding; (iii) frozen treatment cores would release more P during flood incubations than unfrozen control cores; and (iv) processes controlling P release during flood incubations would change after FT due to changes in the primary pools of P in the soil cores. </p> <p> </p> <p>On average, soil cores that underwent FT released greater amounts of P than unfrozen cores over the course of the 3-week experimental flood incubation. Phosphorus release in both unfrozen control and FT treatment cores during flooding was explained in part by soil extractable Al and Fe and redox status; however, P release was influenced by soil Ca-P in the FT cores to a greater extent than unfrozen cores. Phosphorus release in FT cores occurred faster than in control cores with overlying water concentrations peaking 2 weeks after onset of flooding, followed by lower concentrations at 3 weeks. Whereas control cores had some release and uptake early on but then released P throughout the 3-week incubation—supporting the hypothesis that drivers of P release were different after FT. Interactive effects of FT and flooding suggest that concentration gradients between soil pore water and overlying surface water could have enhanced dissolution of the Ca-P pool, highlighting the importance of floodwater chemistry to P dynamics following FT. This study provides an important link between observed winter floodplain P loss and potential drivers of release and retention, which is critical to informing floodplain restoration design and management through all seasons.</p>

Riverine landscapes are complex. More than just a single channel, they comprise a shifting mosaic of hydrogeomorphic patches with varying physical and biological characteristics. These patches are connected by water during flows of varying magnitude and frequency, at a range of spatial and temporal scales. Combined, landscape complexity and hydrological connectivity create biological diversity that in turn maintains the productivity, ecological function, and resilience of these systems. This thesis investigates the ecological importance of spatial heterogeneity and temporal hydrological connectivity in a dryland floodplain river landscape. It focuses on anabranch channels, and uses major carbon sources in these and adjacent landscape patches as indicators of ecological pattern and process. A conceptual model was proposed, describing the potential effects upon the distribution and availability of major carbon sources of: a) a spatial mosaic of hydrogeomorphic patches in the landscape (e.g. anabranches, river channel, and wider floodplain); and b) four primary temporal phases of hydrological connection during flow pulses (disconnection, partial connection, complete connection, and draining). This was then tested by data collected over a three year period from a 16 km reach of the lower Macintyre River (NSW/QLD Australia). Results were examined at multiple spatial scales (patch scale � river channel vs. anabranches vs. floodplain; between individual anabranches; and within anabranches � entry, middle and exit sites). The data indicate that spatial heterogeneity in the lower Macintyre River landscape significantly influences ecological pattern. Carbon quantity was greater in anabranch channels compared to adjacent river channel patches, but not compared to the floodplain; while carbon quality was greater in anabranch channels compared to both adjacent river channel and floodplain patches. Stable isotope analysis indicated that carbon sources that were predominantly found in anabranch channels supported both anabranch and river organisms during a winter disconnection phase. Other carbon sources found in the main river channel and the wider floodplain appeared to play a comparatively minimal role in the food web. Different phases of hydrological connection between anabranch channels and the main river channel were associated with differences in the availability of carbon sources. In the river channel, draining of water from anabranches (the draining phase) was associated with relatively high concentrations of dissolved organic carbon (DOC) and low concentrations of phytoplankton. Conversely, the disconnection phase was associated with relatively low concentrations of DOC and high concentrations of phytoplankton in the river channel. In anabranch channels and their waterbodies, the disconnection and draining phases were associated with high concentrations of both DOC and phytoplankton. Concentrations of these carbon sources were lowest in anabranches during the partial and complete connection phases. Different hydrological connection phases were also associated with changes in trophic status in the aquatic components of the landscape. On the riverbanks, relatively low rates of benthic production and respiration during the complete connection phase were associated with heterotrophy. The remaining phases appeared to be autotrophic. Benthic production on riverbanks was greatest during the disconnection phase, and respiration was greatest during the partial connection phase. In the anabranch channels, rates of production and respiration were similar during the disconnection phase, and were associated with heterotrophy in the anabranch waterbodies. The remaining phases appeared to be autotrophic. Respiration was greatest in anabranches during the disconnection phase, and production was greatest during the draining phase. Both production and respiration were lowest during complete connection. These differences and changes varied according to the landscape patch examined. At a landscape scale, anabranch channels act as both sinks and suppliers of carbon. High rates of sediment deposition facilitate their role as sinks for sediment-associated carbon and other particulate, refractory carbon sources. Simultaneously, anabranch channels supply aquatic carbon sources from their waterbodies, as well as via processes such as inundation-stimulated release of DOC from surface sediments. Modelled data indicated that water resource development reduces the frequency and duration of connection between anabranch channels and the main river channel. This loss of landscape complexity via loss of connectivity with anabranches has the potential to reduce the total availability of carbon sources to the ecosystem, as demonstrated by a modelled 13% reduction in potential dissolved organic carbon release from anabranch sediments. This thesis has demonstrated the importance of spatial heterogeneity in riverine landscapes, by documenting its association with variability in the distribution and quality of primary energy sources for the ecosystem. It has shown that this variability is augmented by different phases of hydrological connectivity over time. Spatial heterogeneity and hydrological connectivity interact to increase the diversity and availability of ecological energy sources across the riverine landscape, at multiple spatial and temporal scales. This has positive implications for the resilience and sustainability of the system. Anabranch channels are particularly important facilitators of these effects in this dryland floodplain river system. Anabranch channels are �intermediate� in terms of spatial placement, temporal hydrological connection, and availability of carbon sources; of high value in terms of high-quality carbon sources; and relatively easy to target for management because of their defined commence-to-flow levels. Further research should be directed toward evaluating other ecological roles of anabranch channels in dryland rivers, thereby providing a more complete understanding of the importance of connectivity between these features and other patches. This knowledge would assist management of floodplain river landscapes at larger regional scales, including amelioration of the effects of water resource development.