Ecosystem functioning on arid and semi-arid floodplains may be described by two alternate traditional paradigms. The pulse-reserve model suggests that rainfall is the main driver of plant growth and subsequent carbon and energy reserve formation in the soil of arid and semi-arid regions. The flood pulse concept suggests that periodic flooding facilitates the two-way transfer of materials between a river and its adjacent floodplain, but focuses mainly on the period when the floodplain is inundated. We compared the effects of both rainfall and flooding on soil moisture and carbon in a semi-arid floodplain to determine the relative importance of each for soil moisture recharge and the generation of a bioavailable organic carbon reserve that can potentially be utilised during the dry phase. Flooding, not rainfall, made a substantial contribution to moisture in the soil profile. Furthermore, the growth of aquatic macrophytes during the wet phase produced at least an order of magnitude more organic material than rainfall-induced pulse-reserve responses during the dry phase, and remained as recognizable soil carbon for years following flood recession. These observations have led us to extend existing paradigms to encompass the reciprocal provisioning of carbon between the wet and dry phases on the floodplain, whereby, in addition to carbon fixed during the dry phase being important for driving biogeochemical transformations upon return of the next wet phase, aquatic macrophyte carbon fixed during the wet phase is recognized as an important source of energy for the dry phase. Reciprocal provisioning presents a conceptual framework on which to formulate questions about the resistance and ecosystem resilience of arid and semi-arid floodplains in the face of threats like climate change and alterations to flood regimes.

1. The leaching rates of filterable reactive phosphorus (FRP) and dissolved organic carbon (DOC) from five leaf litter types commonly occurring in urban environments in Mediterranean regions of Southern Australia were compared. The relative composition, bioavailability and oxygen demand of this DOC were also assessed. Four tree species were assessed, including the native river red gum (Eucalyptus camaldulensis) and three introduced deciduous species, the English elm (Ulmus procera), London plane (Platanus acerifolia) and white poplar (Populus alba). Grass cuttings (mixed species) were selected as a common garden waste. 2. Except for English elm, the majority of FRP and DOC was released within the first 48 h. Grass cuttings released the highest amount of FRP with white poplar releasing the most DOC. Species that released relatively high amounts of DOC (white poplar, English elm, river red gum) released relatively low amounts of FRP. Conversely, species that released relatively low amounts of DOC (grass cuttings and London plane) tended to release relatively high amounts of FRP. 3. Analysis of DOC composition, combined with the differing oxygen demand and DOC depletion curves, demonstrated that there were substantial differences in the DOC leached from the leaf litter of the different species. Biochemical oxygen demand and the biodegradability of the DOC was positively correlated with the proportion of hydrophilic and hydrophobic acids present in the leachate. 4. These results demonstrate that simple measurements of nutrient release per gram of leaf litter would be insufficient to predict the ecological impact on receiving waters resulting from changes in dominant vegetation. Furthermore, the use of traps to prevent particulate leaf material from entering streams may have limited potential for reducing the load of dissolved nutrients. We conclude that any changes to vegetation type which substantially alter the timing of leaf fall or the composition of leaf leachates should be avoided.