eReefs AIMS-CSIRO Aggregations of river tracer model outputs

This generated data set contains summaries (daily, monthly) of the eReefs CSIRO river tracers model v2.0 (https://research.csiro.au/ereefs) outputs at 4km resolution, generated by the AIMS eReefs Platform (https://ereefs.aims.gov.au/ereefs-aims). These summaries are derived from the original daily model outputs available via the National Computing Infrastructure (NCI) (https://dapds00.nci.org.au/thredds/catalogs/fx3/catalog.html), and have been re-gridded from the original curvilinear grid used by the eReefs model into a regular grid so that the data files can be easily loaded into standard GIS software. These summaries are updated in near-real time daily and are made available via a THREDDS server (https://thredds.ereefs.aims.gov.au/thredds/ ) in NetCDF format.

In addition to the variables containing single river data, we have added an 'all_rivers' variable which shows the total river water concentration (%) by combining all river output into a single variable.

The eReefs river tracers model output contains passive river tracer results derived from version 2.0 of the 4km-resolution regional-scale hydrodynamic model of the Great Barrier Reef (GBR4). In the model, tracers are released at the river's mouth into its surface flow. These tracers move with the ocean currents, becoming more dilute as they spread out and mix with the ocean water, allowing the concentration of river water to be tracked over time. These tracers show the fraction of the water, at any given location, associated with each river.

This model configuration and associated results dataset may be referred to as "GBR4_H2p0_Rivers" according to the eReefs simulation naming protocol.

Description of the data:

The data shows the percentage concentration of river water in the marine water. This is a good proxy for the extent of flood plumes associated with the major rivers along the Queensland coastline flowing into the Great Barrier Reef Marine Park. Flood plumes deliver sediments and nutrients into the ocean, both of which can result in detrimental effects on seagrass and reef habitats. Very low salinity concentrations in flood plumes can also cause bleaching and mortality on inshore reefs (this occurred during the flooding on Virago shoal off Townsville after the 2019 floods).This dataset represents only the concentration of river water in the marine environment. It does not model the changes in salinity, the nutrients levels or the sediment concentration in the water. These variables are calculated in the eReefs hydrodynamic model (salinity) and the biogeochemical model (nutrients and sediment). The river tracer is uniquely useful for tracing the origin of flood water back to the source river.

The movement of the river water is driven by the surface ocean currents, that are driven largely by the wind. During most months the south easterly trade winds push the plumes back toward the coast in a northern direction. During the monsoon season, which is strongest between February and March, the winds drop and become more variable in direction. This means that flooding during these months is more variable in direction, occasionally moving southward and out to sea, sometimes reaching the mid shelf reefs. The width of the continental shelf narrows north of Townsville, resulting in it being easier for the flood plumes to reach the mid and outer reefs. Most significant flood plumes occur during the wet season from November to April. Flood plumes are less likely during the dry season from May to October.

The plumes from some of the larger rivers can travel extensive distances during large flooding events. For example during 2019, flood waters from the Burdekin river travelled 700 km north along the coast, reaching Lizard Island. In 2017 the flood waters of the Fitzroy river reached the Whitsundays (450 km north) and the Normanby river water reached the tip of Cape York (440 km north).

The rivers with the biggest discharge resulting in large flood plumes the Burdekin, Herbert, Tully, Johnstone, Russel, Mulgrave, Normanby, Fitzroy and Mary rivers.

The following is a summary of the rivers with significant flood plumes during each year:
2015 Normanby, Mulgrave (minor), Johnstone (minor), Herbert (minor), Fitzroy, Mary
2016 Normanby, Mulgrave (minor), Tully (minor), Burdekin, Fitzroy, Mary (minor)
2017 Normanby, Johnstone, Herbert (minor), Tully (minor), Burdekin (major), Pioneer, Fitzroy (major), Burnett (minor), Mary (minor)
2018 Normanby, Mulgrave, Johnstone, Tully, Herbert, Burdekin, Mary (minor)
2019 Normanby, Daintree, Mulgrave, Johnstone, Tully, Herbert, Haugton (major), Burdekin (major), Pioneer (minor)
2020 Normanby (minor), Burdekin (minor), Fitzroy (minor)
2021 Normanby, Mulgrave (minor), Johnstone (minor), Tully (minor), Herbert, Burdekin (major), Fitzroy, Burnett (minor), Mary (minor)
2022 Normanby (minor), Daintree (minor), Mulgrave (minor), Johnstone (minor), Burdekin, Burnett (minor), Mary (major), Brisbane (minor), Logan (minor)
2023 Normanby, Herbert, Haugton (minor), Burdekin, Fitzroy (minor)

Method:

A description of the processing, especially aggregation and regridding, is available in the "Technical Guide to Derived Products from CSIRO eReefs Models" document (https://nextcloud.eatlas.org.au/apps/sharealias/a/aims-ereefs-platform-t...).

Data Dictionary:

Variables:
- nom: [% river water] Normanby
- mul: [% river water] Mulgrave and Russell
- jon: [% river water] Johnstone
- her: [% river water] Herbert
- bur: [% river water] Burdekin
- fit: [% river water] Fitzroy
- mar: [% river water] Mary
- dai: [% river water] Daintree
- bar: [% river water] Barron
- tul: [% river water] Tully
- hau: [% river water] Haughton
- don: [% river water] Don
- con: [% river water] O'Connell
- pio: [% river water] Pioneer
- bnt: [% river water] Burnett
- fly: [% river water] Fly
- cal: [% river water] Calliope
- boy: [% river water] Boyne
- cab: [% river water] Caboolture
- log: [% river water] Logan
- pin: [% river water] Pine
- bri: [% river water] Brisbane
- all_rivers: [% river water] Aggregation of all river outputs. This is a numerically addition of all single river variables to determine the total river water concentration (%).
- time: [days since 1990-01-01 00:00:00 +10] Time
- zc: [m] Z coordinate (depth) - depth slices
- latitude: [degrees_north] Latitude (geographic projection)
- longitude: [degrees_east] Longitude (geographic projection)

Dimensions:
- time
- k (variable: zc)
- latitude
- longitude

Depths:

This data set contains the following depths, which are a subset of the depths available in the source data set [m]:
-0.5, -1.5, -3.0, -5.55, -8.8, -12.75, -17.75, -23.75, -31.0, -39.5, -49.0, -60.0, -73.0, -88.0, -103.0, -120.0, -145.0.

Limitations:

This dataset is based on a spatial and temporal model and as such is an estimate of the environmental conditions. It is not based on in-water measurements. Furthermore, it should be noted that the river tracer product tracks the concentration of river water. It does not track sediment or nutrient in the water.

As part of research into determining a suitable river concentration threshold for visualisations, we undertook many comparisons between the estimated flood plume extent from eReefs and those visible in Sentinel 2 satellite imagery. From this we found that the plume extent from eReefs was generally accurate to within about 10 km, with the most likely reason for the difference being slight errors in the model due to wind. The strength and direction of the wind is the predominant factor in determining the spread of the flood plumes. As a result any small errors in the modelling of the wind will lead to errors in the flood plume boundaries. The eReefs hydrodynamic model is driven by wind data from the Bureau of Meteorology's Access-R weather model, which is a forecast. It has a resolution of 12 km and so it is surprising that the eReefs model is as spatially accurate as it is. Part of the reason for this is that while the wind occasionally pushes the plumes offshore, the main determinant of the distribution is the dynamics of buoyant plumes. The rotation of the Earth acts to deflect to the left (in the Southern Hemisphere) any relative increase in motion between fluid layers. One such relative motion is a buoyant plume flowing over the top of denser ocean water. Deflected left on a river discharging along an east coast means it being pushed towards the coast. Thus, the plumes are trapped near the coast. The distance to which they spread from the coast is also set by this balance between density driven flow and the Earth’s rotation, something ocean models are very good at.

The eReefs model tracks the percentage river water concentrations to very low levels, such as 1 part per million. At very low concentrations there is likely to not be ecologically relevant. When comparing the plume extents from the river tracer data with flood plumes visible in Sentinel 2 imagery we found that a concentration of 1% river water closely aligned with the visible edge of the flood plumes, where the water is darker and slightly green due to the increased levels of algae in the water.