| Abstract | Floodplain ecosystems support an abundant and diverse range of species. However, in many agricultural landscapes, hydrological and land use changes have caused a decline in the extent and condition of wetlands. There are numerous hydrological based concepts which have helped provide a basis for understanding, predicting and managing the ecological impacts of hydrological changes on floodplain systems. However, while hydrology focused concepts have proved highly informative, they have also lead to much research being exclusively focused on the independent effects of hydrological factors (e.g. reductions in stream flow and flood events). This may limit the ecological understanding and management in three important ways. Firstly, as most studies are carried out in floodplains of large perennial rivers, many current concepts neglect smaller non-flowing habitats, such as floodplain wetlands. Secondly, as research is often exclusively focused on hydrological factors it may not adequately consider the additional impacts of other drivers, such as land use factors. Thirdly, current research gives little consideration to the nature of interactions between different hydrological and land use factors and how they may exacerbate and mitigate effects.
To help test these potential limitations, two alternate hypothesis were developed for the ecology of floodplain wetlands. The “hydrology hypothesis” views these systems as being driven exclusively by hydrological factors. In contrast, the “interactive hydrology-land use hypothesis” considers hydrological and land use factors and their interactions as drivers of ecological patterns in floodplain wetlands. These competing hypotheses are not designed to test which factors are of most relative importance, but to ask whether hydrology focused research is limited by not considering land use and interactions. These hypotheses were tested using data from surveys on the dominant tree species, Eucalyptus camaldulensis Dehnh., of floodplain wetlands in the Condamine Catchment, south east Queensland, Australia. The catchment has undergone extensive hydrological and land use alteration to support agricultural production and therefore offers an ideal setting to test these competing hypotheses.
The hydrological characteristics (inundation frequency, river connectivity, groundwater depth and rain volume) of 102 modified and 149 ‘natural’ unmodified wetlands across the Condamine Catchment were determined using data derived from satellite imagery and digital elevation models and compared. The hydrology of unmodified and modified wetlands differed significantly (ANOVAs; p<0.001), with unmodified wetlands on average, less connected to the river and characterised by significantly lower rain volume.
The condition of the dominant riparian species, E. camaldulensis (as measured by crown vigour), and stag abundance (all trees) in three broad size classes (small trees: <20cm; medium trees: 20-50cm; and large trees: >50cm cbh) at 37 unmodified
ii
wetlands were surveyed and modelled against hydrological metrics and land use factors (grazing, agricultural land cover and remnant vegetation cover) using generalized linear models (GLMs). Specifically, this study tested whether exclusively hydrological models were better than a hydrological plus land use models for understanding variation in crown vigour of E. camaldulensis) and stag abundance. Crown vigour and stag abundance was consistently best explained by hydrological factor only models (GLMs; p<0.05). Consequently, a hydrological-focused perspective (hydrology hypothesis) may not necessarily be limited when studying aspects of tree condition, such as crown vigour, in floodplain wetlands.
The distribution (occurrence) of E. camaldulensis in five size classes (<10, 10-20, 20-50, 50-75 and >75cm cbh) across 37 unmodified wetlands was also modelled against hydrology and land use factors using GLMs. Contrasting, with models on crown vigour and stag abundance, E. camaldulensis occurrence was significantly related to both hydrological and land use factors (GLMs; p<0.05). Models which included both hydrological (distance from weir, river connectivity and groundwater depth) and land use factors (agricultural land cover and grazing intensity) performed better (R2 0.04 to 0.24 and AUC 0.06 to 0.16 greater) than those developed using only hydrological factors. It was concluded that hydrology focused research that does not consider land use may not be suitable for understanding impacts on the distribution of E. camaldulensis.
A Bayesian network modelling approach was used to integrate the results from the individual studies to develop a broad model of the drivers of E. camaldulensis occurrence and condition and overall wetland condition and to explore possible interactions between these drivers. E. camaldulensis response to hydrology and land use factors was better characterised by their combined interactive effects than their independent effects. Interactions identified were classified as synergistic, antagonistic and qualitative. Of these, qualitative interactions (an interaction between two factors that causes a change in both the magnitude and direction of response) have not been described in previous ecological research and as such, may be important for broader thinking about interactions in ecosystems. It was concluded that the failure to consider interactions, and how they vary (e.g. synergistic, antagonistic and qualitative), may lead to an over or under estimation of how species relate to their environment and potentially counterproductive management actions.
The conflicting responses of the variables tested suggest that neither the hydrology hypothesis nor interactive hydrology – land use hypothesis was universally applicable for understanding all aspects of E. camaldulensis in the floodplain wetlands examined. The two hypothesises are therefore not mutually exclusive; both are applicable depending on the aspect examined (i.e. crown vigour or occurrence). As a consequence, neither can be rejected until more research is carried out on a wider range of hydrological and land use factors. Nonetheless, the results did show that it is not valid to assume that only hydrological factors are important drivers of all
iii
ecological aspects in these systems. Consequently, the additional consideration of land use factors is needed to give a more complete understanding of how this species relates to its environment. Without this broader examination, then other factors (e.g. land use) limiting the occurrence of E. camaldulensis, as well as how it responds to interactions between these factors, may be overlooked and as such ecological understanding and management may be missing important information and thus could be ineffectual (or even detrimental) for wetlands.
The results of this research suggest that management of floodplain systems, particularly the vegetation of floodplain wetlands, which focuses exclusively on hydrological factors and does not consider land use and different types of interactions, may be significantly limited. Accepting the hydrology hypothesis and a hydrological focus for wetlands in the Condamine Catchment would mean that negative land use impacts from grazing and agricultural may be overlooked and as such, ecological conservation measures limited. Additionally, it would mean that the response of E. camaldulensis to multiple hydrological (e.g. groundwater, river connectivity, inundation frequency) and land use (e.g. grazing) factors could be over or under estimated if they interact synergistically, antagonistically and / or qualitatively. The results of this thesis therefore highlight some potentially significant limitations of hydrology focused research and management which currently predominates in floodplain systems. Future research should (1) more broadly test the limitations of the hydrology and interactive hydrology-land use hypotheses in floodplain systems; (2) test the importance of considering different types of interactions, especially qualitative interactions, for a range of different biota in different settings to examine whether they apply more broadly to other species and ecosystems. |
|---|