Wetlands and Floodplains: Connectivity and Hydroecological Function

Part I - The role of overbank floods in transporting sediments and nutrients to the Great Barrier Reef lagoon

J. Wallace, A. Hawdon, R. Keen, F. Karim, L. Stewart and J. Kemei
CSIRO Land and Water, Davies Laboratory, Townsville

Part II - Quantification of overbank and channelised wetland connectivity in the Tully-Murray floodplain

F. Karim¹, J. Wallace¹, A. Kinsey-Henderson¹, A. Hawdon¹, R. Keen¹, A. H. Arthington², P. Godfrey² and R. G. Pearson^3
1 CSIRO Land and Water, Davies Laboratory, Townsville
² Australian Rivers Institute, Griffith University, Nathan
³ School of Marine and Tropical Biology, James Cook University, Townsville

MTSRF Project 3.7.4 - Wetlands and floodplains:  Connectivity and hydro-ecological function

Extracts

Part I Executive Summary (Wallace et al.)

The potential for agricultural practices to enhance the loads of sediment and nutrients entering the Great Barrier Reef (GBR) lagoon has led to the development of a number of catchment based 'Water Quality Improvement Plans' (WQIPs). These plans identify the current constituent loads along with a set of management practices to reduce them. The current sources and annual average loads of sediment and nutrient are estimated using gauged river flow and concentration data and runoff water quality modelling. However, in catchments that are subject to frequent flooding, standard river gauges may significantly underestimate overbank flood flows. This is the case in the Tully and Murray catchments where the river gauges do not record the total catchment discharge during floods very well. For example, during the thirteen flood events between 2006 and 2008 investigated in this study, the Tully river gauge at Euramo recorded only 36-88% of the flood discharge, while the Upper Murray gauge recorded only 11-27% of the flood discharge. Furthermore, current ocean sediment and nutrient loads are based on concentrations measured within the rivers, yet it is not known what the sediment and nutrient concentrations are in overbank flood waters.

This report addresses these issues by updating current estimates of sediment and nutrient loads from the Tully-Murray floodplain to the GBR lagoon, taking explicit account of flood events. New estimates of flood discharge that include overbank flows are combined with direct measurements of sediment and nutrient concentrations in flood waters to calculate the loads of sediment and nutrient delivered to the ocean during the above mentioned flood events between 2006 and 2008. Although absolute concentrations of sediment and nutrient were quite low, the large volume of water discharged during floods means that they make a large contribution (30-50%) to the marine load. By not accounting for flood flows correctly, previous estimates of the annual average discharge are 15% too low, and annual loads of nitrogen and phosphorus are 47% and 32% too low respectively. However, as sediments may be source limited, accounting for flood flows simply dilutes their concentration and the resulting annual average load is similar to that previously estimated.

A second important feature emerges from the flood water quality data, which show that flood waters carry more dissolved organic nitrogen (DON) than dissolved inorganic nitrogen (DIN) and this is the opposite of their concentrations in river water. Consequently, DON loads to the ocean may be around twice those previously estimated from riverine data. Whereas the main source of DIN is from agricultural land, the main source of DON is likely to be from upper catchment rainforest. If runoff to the GBR lagoon has increased due to land drainage, there may therefore be an enhanced and biologically available DON load to the ocean arising from the upper catchment rainforest.

The implications of the flood water quality studies in the Tully and Murray catchments, and potentially for other GBR catchment WQIPs, are as follows:

1.    Overbank floods can make a large contribution to the marine load of sediment and nutrients and much of this load may not be recorded by standard river gauges.

2.    In GBR catchments where floods are a significant proportion of the annual flow, current marine load estimates of sediment and nutrients (based on gauged flows, measured river concentrations and modelling) are probably too low, by significant amounts, depending on estimation method and constituent.

3.    The size of this underestimate in any year will depend on the number and size of overbank flood events in that year. This will make the monitoring of any underlying trends in ocean loads difficult unless it is possible to remove inter-annual variability.

4.    Monitoring of marine loads will take a significant number of samples of both river and flood flows (in time and space) – otherwise the large uncertainties in mean loads may be misleading and it may be difficult to detect any load reduction trends.

5.    The cause of the above underestimate in loads is mainly due to the poor recording of flood (overbank) discharges by river gauges, but also to differences in flood water and river water quality concentrations. 

6.    Flood waters can carry more DON than DIN, and this is the opposite of their concentrations in river water. Consequently DON loads to the ocean may be much higher than those previously estimated from riverine data.

7.    WQIP actions that focus on farm interventions in agriculture will potentially reduce DIN loads.

8.    Reductions in DON (and sediment) loads that arise outside the floodplain require different interventions to those used in agriculture to reduce DIN, e.g. measures that slow and reduce drainage and the introduction and/or rehabilitation of riparian zones and wetlands.

The inaugural flood water quality data collected in the Tully and Murray catchments has demonstrated the importance of obtaining observations from the key processes that control the marine loads that are of concern. In the Wet Tropics catchments studied, in addition to chanelised flow, overbank flooding is a primary material transport mechanism and it is very difficult to adequately capture this process in monitoring and/or modelling schemes that are entirely river based. There is therefore a clear need to obtain estimates of the contribution that floods make to marine loads in other GBR catchments.

Part II Executive Summary (Karim et al.)

Hydrological connectivity between floodplain wetlands and rivers is the principal driving mechanism for the diversity, productivity and interactions of the major biota in river-floodplain systems. This wetland connectivity is initiated by overbank floods, but it can continue after flooding via the stream and rain network on the floodplain. This report describes the application of hydrodynamic modelling to quantify the timing, duration and spatial extent of both flood-induced overbank connectivity and post-flood drainage network connectivity between a number of wetlands and the main rivers in the Tully-Murray catchment, northern Queensland, Australia.

The wetlands on the floodplain were identified using high resolution laser altimetry (LiDAR) data incorporated with aerial photogrammetry data to form a digital elevation model (DEM) of the floodplain. Propagation of flood waves and associated floodplain inundation were simulated using a 2-D hydrodynamic model (MIKE 21) that computed water depth and flow velocity on a 30 m grid. Connectivity between ten wetlands of different types (natural and artificial) and the two main rivers (Tully and Murray) was estimated for flood events of one, twenty and fifty year recurrence intervals. The duration of connection of individual wetlands varied from 0 to 12 days depending on flood magnitude and location in the floodplain, with some wetlands only connected during large floods. All of the wetlands studied were connected to the Tully River for shorter periods than they were to the Murray River, due to their proximity to the Murray River and the higher bank heights and levees on the Tully River. These variations in wetland connectivity could affect the movement of aquatic biota during floods and the variability of habitat and biodiversity of individual wetlands.

Post-flood wetland connectivity via the drainage network was quantified using a 1-D hydrodynamic model (MIKE 11) to calculate the timing and duration of connectivity of seven wetlands of different types (natural and artificial) and the two main rivers during 2007 and 2008. The location and size of the wetlands and the extent and size of the stream and drain network were identified using high resolution laser altimetry (LiDAR) and these data formed key inputs to the hydrodynamic model. The MIKE 11 model was calibrated using measured discharges and water depths at several locations in the rivers, streams and drains. We found that wetlands which are located near the rivers and/or have good network connection maintain longer connection times with the rivers. Drainage network connectivity to both rivers varied from 30 to 365 days, and was much greater than flood inundation connectivity for the same wetlands (0-12 days) described above. The connectivity of artificial wetlands varied greatly, from ten to one hundred percent of the year, according to the type of network connection they have; a result that has important implications for the location of these types of wetlands. We also show how this kind of connectivity modelling can be used to identify when water levels in a drainage network fall below critical thresholds for fish movement using readily available river gauge data. These types of relationships are central to the concept of setting environmentally acceptable flows in floodplain rivers. Quantitative connectivity modelling will also be useful for helping to explain spatial variation in habitat structure, water quality and the composition of biotic communities in individual wetlands over time.