Using coupled models to place constraints on fluvial input into , Phaedra Upton1, Rachel Skudder2 and the Lakes Team3 [email protected], 1GNS Science, 2Victoria University of Wellington, 3GNS Science + Otago University + Victoria University of Wellington

2004: no large storms 0 2000 kg/sec 1/1/10

daily rainfall Godley River modelled suspended sediment load

Tasman River

Lake Tekapo 1/1/00

0 200 mm

Lake Pukaki 1995: large summer storm 0 2000 kg/sec

Hopkins River Box core #1

1/1/90 daily rainfall

modelled suspended Lake Ohau Box core #2 Raymond Film Services sediment load

Ahuriri River Figure 2: following a large rainfall event in its catchment. A sediment laden inflow plunges into the lake and leaves the surface waters clear. We use HydroTrend (Kettner and Syvitski 2008, Computers and 6 m core Daily Rainfall (mm) 0 200 mm Geoscience, 34) to calculate the overall sediment influx into the lake and 1/1/80 couple it to a conceptual model of how this sediment might be distributed 1969: large winter storm through the lake basin depending on the season to produce model cores. 0 2000 kg/sec 0 3 6 12 18 24km

Figure 1: Located east of the main divide in the central Southern Alps, the Mackenzie Lakes; Ohau, Pukaki and Tekapo, occupy fault Figure 3: Map of Lake Ohau showing the location of the three cores we controlled glacial valleys and contain high resolution sedimentary compare our models to. daily rainfall records of the last ~17 ka. These sediments potentially contain a 1/1/70 modelled suspended record of climatic events and transitions, earthquakes along the Alpine sediment load Fault to the northwest, landscape response during and following deglaciation and recent human-influenced land use changes. Our study focuses on Lake Ohau, the smallest and deepest of the three 0 200 mm lakes.

Delta Deep Distal 1/1/60 1/1/10 Dec 2010 0 100 200 300 mm 1 May 2010 1 May 2010 Figure 4: Daily rainfall (NIWA) from 1960 - 2011 averaged across the Ohau catchment and used as an input for HydroTrend models. Three characteristic years are shown. Rainfall (blue) and suspended sediment load calculated using HydroTrend (red). 1 May 2000 1 May 2000 1/1/00

Dec 1995 Modelled Figure 6: Upper 25 cm of Lake Ohau Distal Core sediment core 6M-1. Left panel: line scan image with RGB data overlain. Age model is Jan 1994 Dec 2010 Core OHAU-6M-1 based on cuplet counts with red age based on 1 May 1990 1 May 1990 Spectral (RGB) from Lake Ohau (cumecs) Cs dating (Ditchburn et al. 2011, GSA Abstract). Line scan Age 1/1/90 Depth 00 00 00 and spectral Model 00 Middle panel: zoom into 1960 - 2000, RGB (cm) 0 0 0 0 3 2 4 1 16 12 (RGB) (years) 14 10 specta with precipitation anomalies (Ummenhofer and England 2007, DRY Dec 1984 Journal of Climate, 20). Total monthly outflow 5 WET 1995 Dec 1995 00 from Lake Ohau (cumecs). Blue and orange DRY Mar 1982 95 Jan 1994 correlation lines highlight wet and dry years. 10 90 Right panel: modelled distal core with summer June 1981 1990 1/1/80 85 layers shown as a massive brown deposit and Dec 1979 1 May 1980 1 May 1980 15 80 WET winter layers shown by greyscale. DRY May 1978 75 1985 Dec 1984

20 70 WET Correlation of individual layers with individual DRY storm events is difficult as the size of the 1964 1980 60 resultant layer likely depends on precipitation 25 WET Dec 1979 and other factors such as the time since the last DRY 1/1/70 Dec 1969 major storm (e.g., McCulloch et al. 2003, Nature Sept 1969 1975 May 1978 421). For example, both 1994 and 1995 had DRY 1 May 1970 1 May 1970 very large storms, the 1995 storm being the WET DRY largest, yet only one large layer is obvious in the 1970 core. Prior to the 1994, the previous large storm Sept 1969 Massive layer from stratified lake was in 1984. Thus between 1984 and 1994, 10 DRY years of material was available to be moved Layers from mixed lake with large storm 1965 1/1/60 through the catchment. 0 1000 2000 Massive layer from stratified lake Daily Suspended Sediment Layers from mixed lake with no large storm 1960 Load (kg/s) 1 May 1960

Figure 5: Model results from HydroTrend shown as daily suspended sediment load and converted to synthetic cores for three sites within the lake. On the left hand side of each core, spring/summer inflows are assumed to be mixed through the season producing a massive layer, coloured brown. Winter events form individual layers with the size of the inflow proportional to the greyscale, dark = small, light = large. The biggest winter inflows show up as white layers. On the right hand side of each core, spring/summer inflow events also form individual layers and again the greyscale is proportional to the size of the inflow. Tie lines between the delta and deep regions are shown every year at the beginning of May. Tie lines between the deep and distal regions are shown every 5 years, also at the beginning of May.

Density of lake water and river inflows from 1960 to 2012 1002 Conceptual model of sediment inflow into Lake Ohau ) rainfall and sediment input -3 1001 Figure 7: Conceptual model of Lake Ohau Figure 8: Calculated density of lake showing mixed (generally winter, April - Novem- density of inflow defined by temperature of underflows Mixed whole lake 1000 ber) and stratified (generally summer, December the water and concentration of sediment water (mixed in cooler months, stratified in Density (kg m Stratified base - March) time periods. During the warmer interflows warmer months) and river inflows as a 999 Stratified top function of temperature and sediment con- months, the lake is stratified and sediment interflows into stratified lake plumes are expected to form interflows with sedi- centration. When the lake is mixed, all ment being deposited along the entire lake basin. 1960 1970 1980 1990 2000 2010 flows plunge to the bottom of the lake. underflows into mixed lake When the lake is stratified, most flows are During the cooler months, the lake is mixed and Delta sediment inflows are expected to plunge to the 1001 interflows but highly sediment laden flows Deep ) are dense enough to plunge to the bottom bottom of the lake. Sediment deposition occurs -3 1000.5 predominately on the delta and in the deep lake stratified of the lake. Distal stratified underflows with the distal regions receiving minimal winter mixed 1000 Mixed whole lake mixed 999.5 Stratified base Density (kg m interflows

Thick/Coarse Thin/Fine 999 Stratified top Winter Winter Summer Summer 1980 1985 1990

Box core #1 Box core #2

summer 10/11

summer 10/11

Sedflux model of Lake Ohau: 1960-2012 Grain size Core site #1 Core site #2 0 0 0 28/12/2010 28/12/2010

20 0.01

40 23/12/1995 23/12/1995

60 0.02 0.05 Depth (m) 80 Core site #1 Core site #2 0.03 100 ?1995 Burial Depth (m) large storm Burial Depth (m) ?local debri flow 120 12/5/1979 from side valley? 0.04 12/5/1979 0 2 4 6 8 10 12 14 16 poss during 1995 Y-Distance (km) 0.1 8/9/1969 large storm 8/9/1969 0.05 3 4 5 6 7 8

Grain size (φ) 7.5 7.55 7.6 7.65 8.1 8.15 8.2 8.25 Figure 9: Sedflux model of Lake Ohau run over 53 years (1960-2012) using HydroTrend output as input into the Grain size (φ) Grain size (φ) model. Bedload is deposited within 1 im of the delta.

Comparison of models and cores:

The models greatly simplify the along lake sediment distribution. We use our conceptual model to convert the HydroTrend output into synthetic cores. Temperature records from the lake, short cores and the models suggest that the delta region of the lake records individual events while the depocentre and distal regions of the lake record layers which represent either a stratified or a mixed lake. The Sedflux model underestimates the sediment reaching the distal end of the lake. Is this because Sedflux does not include statification of the water column and thus all sediment laden flows are modelled as underflows? The grainsize distribution from HydroTrend into Sedflux are very simplified, i.e., there is no change in grainsize distribution with an increase in discharge.