Impact model: FLake-BU

Sector
Lakes (global)
Region
global

FLake is a one-dimensional bulk lake model that represents the temperature profile with a two-layer parametric (self-similar) structure: a mixed layer and a thermocline that extends to the lake bottom. It includes a thermodynamic ice module and parameterizations of the surface fluxes, and was designed to represent lakes in numerical weather prediction and climate models. FLake-BU applies FLake without calibration to the ISIMIP global set of representative lakes (one lake per 0.5° grid cell).

Information for the model FLake-BU is provided for the simulation rounds shown in the tabs below. Click on the appropriate tab to get the information for the simulation round you are interested in.

Person responsible for model simulations in this simulation round
Yan Tong: y.tong@bangor.ac.uk, 0000-0003-2035-8360, Bangor University (UK)
Additional persons involved: R Iestyn Woolway
Output Data
Experiments: (*) esm-picontrol_2021soc_default, esm-scen7-vl_2021soc_default, esm-hist_2021soc_default, esm-scen7-h_2021soc_default
Climate Drivers: EC-EARTH3-ESM-1-1, UKESM1-3-LL
Date: 2026-11-20
Basic information
Model Version: FLake
Model Output License: CC0
Model License: MIT License
Simulation Round Specific Description: FLake-BU applies the FLake model (Mironov 2008) without calibration. Following FLake's model assumptions, each lake is simulated with its mean depth, limited to 1–50 m, rather than its maximum depth, so the simulated lakes are shallower than in models that resolve the lake bathymetry; for lakes deeper than 50 m, bottemp refers to a 50 m "false bottom". The light extinction coefficient is derived from the mean depth (Håkanson 1995) and is constant in time. Snow on lake ice and the bottom-sediment module are switched off; shortwave radiation reaching the lake bottom is returned to the water as heat, so the lake energy budget closes. The model runs at a 1-hour time step with daily-mean forcing.
Reference Paper: Main Reference: Mironov, D. V. et al. Parameterization of lakes in numerical weather prediction. Description of a lake model.COSMO Technical Report No. 11, Deutscher Wetterdienst, Offenbach am Main, Germany.,,,2008
Resolution
Spatial aggregation: regular grid
Horizontal resolution: 30arcmin
Vertically resolved: Yes
Number of vertical layers: 10 layers. Water temperature is reported as the mean over each of 10 equal-thickness layers spanning the lake surface to the model lake depth. Layer means are calculated from FLake’s two-layer temperature profile, which uses a self-similar shape function for the thermocline. The depth variable gives the layer centres, and depth_bnds gives the layer boundaries.
Additional spatial aggregation & resolution information: FLake simulations are conducted for the representative lake of each 0.5x0.5 resolution grid cell. Each lake is forced with the climate of its grid cell.
Temporal resolution of input data: climate variables: daily
Additional temporal resolution information: FLake-BU is integrated with a 1-hour time step (24 steps per day), and daily means of all output variables are reported. The sub-daily step is needed for numerical stability.
Input data
Simulated atmospheric climate data sets: EC-Earth3-ESM-1-1, UKESM1-3-LL
Other data sets: Lakes static information
Spin-up
Was a spin-up performed?: Yes
Spin-up design: For each GCM, the historical and pre-industrial control simulations shared a 30-year spin-up based on three repetitions of the 1841–1850 esm-picontrol forcing. Both simulations began in 1850 from the same final spin-up state. The H and VL simulations began on 1 January 2022 from the corresponding historical simulation’s final state.
Natural Vegetation
Soil layers: No sediment layers. The FLake-BU bottom-sediment module is switched off.
Calibration
Was the model calibrated?: No
Vegetation
Is co2 fertilisation accounted for?: No
Methods
Snow melt: Not simulated. Snow on lake ice is not represented in FLake-BU; its effect on the surface energy balance enters only through the ice albedo.
Additional questions 1
How did you initialise you lake temperature profile?: Please see the answer in the spin-up section.
How did you set lake depth?: Mean depth of the ISIMIP3 representative lake (HydroLAKES/GLOBathy), limited to 1-50 m; where the mean depth is missing, volume divided by area.
How did you set water transparency?: Light extinction coefficient is calculated from the mean depth D following Håkanson (1995): Kd = 1.1925 · max(D, 1 m)^-0.424 (m-1), with D not limited to 50 m, and Kd limited to 0.07-5 m-1.
Person responsible for model simulations in this simulation round
Yan Tong: y.tong@bangor.ac.uk, 0000-0003-2035-8360, Bangor University (UK)
Additional persons involved: R Iestyn Woolway
Output Data
Experiments: (*) obsclim_2021soc_default
Climate Drivers: GSWP3-ERA5
Date: 2026-11-19
Basic information
Model Version: FLake
Model Output License: CC0
Model License: MIT License
Simulation Round Specific Description: FLake-BU applies the FLake model (Mironov 2008) without calibration. Following FLake's model assumptions, each lake is simulated with its mean depth, limited to 1–50 m, rather than its maximum depth, so the simulated lakes are shallower than in models that resolve the lake bathymetry; for lakes deeper than 50 m, bottemp refers to a 50 m "false bottom". The light extinction coefficient is derived from the mean depth (Håkanson 1995) and is constant in time. Snow on lake ice and the bottom-sediment module are switched off; shortwave radiation reaching the lake bottom is returned to the water as heat, so the lake energy budget closes. The model runs at a 1-hour time step with daily-mean forcing.
Reference Paper: Main Reference: Mironov, D. V. et al. Parameterization of lakes in numerical weather prediction. Description of a lake model.COSMO Technical Report No. 11, Deutscher Wetterdienst, Offenbach am Main, Germany.,,,2008
Resolution
Spatial aggregation: regular grid
Horizontal resolution: 30arcmin
Vertically resolved: Yes
Number of vertical layers: 10 layers. Water temperature is reported as the mean over each of 10 equal-thickness layers spanning the lake surface to the model lake depth. Layer means are calculated from FLake’s two-layer temperature profile, which uses a self-similar shape function for the thermocline. The depth variable gives the layer centres, and depth_bnds gives the layer boundaries.
Additional spatial aggregation & resolution information: FLake simulations are conducted for the representative lake of each 0.5x0.5 resolution grid cell. Each lake is forced with the climate of its grid cell.
Temporal resolution of input data: climate variables: daily
Additional temporal resolution information: FLake-BU is integrated with a 1-hour time step (24 steps per day), and daily means of all output variables are reported. The sub-daily step is needed for numerical stability.
Input data
Observed atmospheric climate data sets: GSWP3-ERA5
Other data sets: Lakes static information
Climate variables: huss, ps, rlds, rsds, sfcWind, tas
Spin-up
Was a spin-up performed?: Yes
Spin-up design: FLake-BU was spun up for 100 years by repeating the 1901–1925 GSWP3–ERA5 obsclim forcing four times.
Natural Vegetation
Soil layers: No sediment layers. The FLake bottom-sediment module is switched off.
Calibration
Was the model calibrated?: No
Vegetation
Is co2 fertilisation accounted for?: No
Methods
Snow melt: Not simulated. Snow on lake ice is not represented in FLake-BU; its effect on the surface energy balance enters only through the ice albedo.
Additional questions 1
How did you initialise you lake temperature profile?: Please see the answer in the spin-up section.
How did you set lake depth?: Mean depth of the ISIMIP3 representative lake (HydroLAKES/GLOBathy), limited to 1-50 m; where the mean depth is missing, volume divided by area.
How did you set water transparency?: Light extinction coefficient is calculated from the mean depth D following Håkanson (1995): Kd = 1.1925 · max(D, 1 m)^-0.424 (m-1), with D not limited to 50 m, and Kd limited to 0.07-5 m-1.