Tropical Reservoir Water Quality Workbook (GLM-AED)

ENVT3362 · GLM 3.3.0 with AED

Overview

In this workbook, we build up the water quality model (AED) for the Tropical Reservoir step-by-step. We start with a model with dissolved oxygen, then we add nutrients, one algal group, and three competing algal groups. We then explore how phytoplankton dynamics respond to different nutrient loads, and finally (optional) add zooplankton grazing. You will set up each case in its own folder, starting from a copy of the Tropical Reservoir model.

Learning objectives

By the end of this workbook, you should be able to:

1. Build your own AED model configuration by switching modules on and off.
2. Explain the feedbacks between dissolved oxygen, nutrients, phytoplankton and zooplankton.

Study site

The reservoir is in the tropics of northern Australia (12.85° S, 131.1° E). It is about 14 m deep at the start of the run, with a crest at 38.17 m AHD. The simulation runs from 15 Nov 2011 to 1 Jan 2014 at an hourly time step. Inflow arrives mostly as wet-season pulses, about 86 million m³ per year.

Cases

Each case has its own folder in glm_casestudies, next to the GLM folder. You make each folder by copying the folder of an earlier case and renaming it, so you never overwrite a case you have already run.

Units. AED reports concentrations in mmol/m³. Phytoplankton biomass is in mmol C/m³. As a rough guide, chlorophyll-a (µg/L) ≈ mmol C/m³ × 12 / 50, assuming a C:Chl ratio of 50.

Setting up

These steps create the first case folder, TropicalReservoir_0_Base. Every later case is a copy of an earlier one, so the changes you make here carry through to all cases.

Step 1 — Create the TropicalReservoir_0_Base folder

Open the glm_casestudies folder that you downloaded last week. Copy the TropicalReservoir folder, paste it into glm_casestudies, and rename the copy TropicalReservoir_0_Base. Open the new folder and delete the output folder inside it; GLM creates a new one the first time you run the case. Your folders should now look like this:

glm_casestudies/
│── GLM/
│── TropicalReservoir/            (original — do not change)
└── TropicalReservoir_0_Base/
    │── aed/          aed.nml, aed_phyto_pars.nml, aed_zoop_pars.nml
    │── bcs/          met.csv, inflow_WQ.csv, outflow.csv
    │── glm3.nml
    │── plots_aed.nml
    │── run_glm.bat
    └── run_glm.sh

Every case folder sits next to the GLM folder, in the same place as TropicalReservoir, so run_glm.bat and run_glm.sh work without any changes.

Step 2 — Set the surface and bottom outputs

Open glm3.nml and find the &output block. Replace the depth-specific output lines so that GLM writes two files: WQ_0.csv at the surface (0.5 m below the water surface, following the water level as it rises and falls) and WQ_1.csv at the bottom (1 m above the sediment):

   ! Depth specific outputs
   csv_point_nlevs   = 2
   csv_point_fname   = 'WQ_'
   csv_point_at      = 1, 0.5
   csv_point_frombot = .true., .false.   ! 1 m from the bottom, 0.5 m from the surface
   csv_point_nvars   = 2
   csv_point_vars    = 'temp','OXY_oxy'

In each later case you only change csv_point_nvars and csv_point_vars, to add the new variables.

Step 3 — Switch off the retention-time tracer

Open aed/aed.nml, find the &aed_tracer block and set retention_time to .false.:

&aed_tracer
    retention_time = .false.
    num_tracers = 1
    ...
/

Why? In GLM 3.3, the inflow columns listed in inflow_vars are passed to the AED variables in order, starting from the second AED variable. Keeping one tracer (aed_tracer) as the first module fills that first place, so oxygen and the nutrients receive their correct inflow concentrations. The tracer is a passive dye and does not affect the results.

Running the model

Run each case from its own folder, in the same way as in the previous lab.

• Windows: double-click run_glm.bat.
• Mac: open Terminal in the case folder (right-click the folder in Finder → New Terminal at Folder), run bash run_glm.sh and choose option 2 (Run with plots_aed.nml). If macOS blocks the GLM app the first time, allow it in System Settings → Privacy & Security and run again.
• Linux: with GLM installed so that the glm command works, run glm --xdisp plots_aed.nml in the case folder.

Each run writes to the output/ folder of that case: output.nc, lake.csv, WQ_0.csv (surface) and WQ_1.csv (bottom). Open the two WQ files in Excel or R to make your plots. Every run overwrites these files.

The plot window shows the variables listed in plots_aed.nml. In each case you replace the &plots block of this file so that it lists only the variables that exist in that case.

Case 0 — Base: Dissolved oxygen (TropicalReservoir_0_Base)

The base case simulates dissolved oxygen only. Surface oxygen rises and falls with the seasons, while the bottom water has almost no oxygen (below 10 mmol/m³) on 638 of the 778 simulated days.

Oxygen enters the reservoir at the surface, from the air and with the inflow. At the bottom, the sediment continuously uses oxygen (the sediment oxygen demand, Fsed_oxy). When the reservoir is stratified, the warm surface layer sits on top of the cooler bottom layer and the two do not mix. The bottom water is then cut off from its oxygen supply, so the sediment draws it down to near zero.

First, let's check that AED is switched on. In glm3.nml, the &wq_setup block must be uncommented and point to the AED namelist:

&wq_setup
   wq_lib            = 'aed'
   wq_nml_file       = 'aed/aed.nml'
   ode_method        = 1
   split_factor      = 1
   bioshade_feedback = .true.
   repair_state      = .true.
/

The oxygen module is set up in the &aed_oxygen block:

&aed_oxygen
   oxy_initial   = 225.0     ! starting concentration (mmol/m3)
   Fsed_oxy      = -20.0     ! sediment oxygen demand (mmol/m2/day)
   Ksed_oxy      =  50.0     ! half-saturation for the sediment demand (mmol/m3)
   theta_sed_oxy =   1.08    ! temperature multiplier
/

Case 1 — Add nutrients, organic matter and silica (TropicalReservoir_1_Nutrients)

With nutrients and organic matter on, but no algae to use them, surface phosphate (FRP) keeps rising from 0.02 to about 0.15 mmol/m³ by the end of the run. Ammonium and nitrate average about 4.4 mmol/m³ each. Mean surface oxygen falls from 110 to 81 mmol/m³, because the breakdown of organic matter uses oxygen.

The new modules add:

The organic matter module links to the nutrient modules through its target variables. For example, don_miner_product_variable = 'NIT_amm' tells AED that mineralised organic nitrogen becomes ammonium. This is why aed_organic_matter is listed after nitrogen and phosphorus.

Case 2 — Add diatoms (TropicalReservoir_2_Diatom)

On their own, diatoms barely grow in this reservoir. Their biomass stays near 0.04 mmol C/m³ and peaks at only 0.8 mmol C/m³ in March 2012. The water is warmer than the diatom optimum on 88% of days, and there is very little phosphate.

The other lines of the &aed_phytoplankton block connect the phytoplankton to the other modules:

  p1_uptake_target_variable   ='PHS_frp'
  n1_uptake_target_variable   ='NIT_nit'
  n2_uptake_target_variable   ='NIT_amm'
  si_uptake_target_variable   ='SIL_rsi'

The *_target_variable lines link phytoplankton dynamics to variables from other modules. Algae use FRP, nitrate, and ammonium as nutrients for growth. When they release organic material or die, the carbon, nitrogen, and phosphorus they contain enter the organic matter pools (OGM_doc, OGM_pon, and others).

Open aed_phyto_pars.nml and look at column 7 ('diatom'). The most important parameters are:

Case 3 — Three phyto groups (TropicalReservoir_3_ThreePhyto)

Under competition, diatoms are largely outcompeted by green algae and cyanobacteria. The overall phytoplankton population increases, which in turn draws down surface phosphate concentrations. As phosphate becomes scarce, phosphorus limitation begins to constrain phytoplankton growth.

settling = 4 lets cyanobacteria change their density between min_rho and max_rho (900 and 1200 kg/m³, set at the end of the &aed_phytoplankton block). This means they can float up to the light, like real Microcystis or Dolichospermum.

Compare the three groups in aed_phyto_pars.nml (columns 3, 7 and 8):

Case 4 — Nutrient load scenarios (TropicalReservoir_4_HighNutrientLoad)

In this case we increase the nutrient load coming from the catchment and look at how the nutrient concentrations in the reservoir and the phytoplankton respond. The baseline for comparison is Case 3.

The mean inflow concentrations become:

Case 5 — Add zooplankton optional (TropicalReservoir_5_Zooplankton)

This case is optional. It adds one zooplankton group to Case 3, grazing on the three phytoplankton groups. With the supplied parameters, the zooplankton cannot sustain itself: it falls from 0.5 to its minimum of 0.01 mmol C/m³ within the first weeks.

The &aed_zooplankton block chooses group 1 from the zooplankton database and sends its waste to the organic matter pools:

&aed_zooplankton
   num_zoops = 1
   the_zoops = 1
   dn_target_variable = 'OGM_don'
   pn_target_variable = 'OGM_pon'
   dp_target_variable = 'OGM_dop'
   pp_target_variable = 'OGM_pop'
   dc_target_variable = 'OGM_doc'
   pc_target_variable = 'OGM_poc'
   dbase='aed/aed_zoop_pars.nml'
/

In aed_zoop_pars.nml, column 1 (zoo01) grazes on three prey, with a preference for each:

  zoop_param%num_prey          = 3
  zoop_param%prey(1)%zoop_prey = 'PHY_green',   Pzoo_prey = 0.4
  zoop_param%prey(2)%zoop_prey = 'PHY_diatom',  Pzoo_prey = 0.1
  zoop_param%prey(3)%zoop_prey = 'PHY_cyano',   Pzoo_prey = 0.5

Its maximum grazing rate (Rgrz_zoo = 0.1/day) is too low to cover its respiration (0.08/day) and mortality (0.01/day). Group 2 (zoo02, Rgrz_zoo = 1.67/day, which does not eat cyanobacteria) can be tried by setting the_zoops = 2.