01
Climate & Earth Sciences
General circulation models contain parameters such as cloud microphysics coefficients and aerosol forcing terms that are difficult to constrain from observations alone. Literature-informed priors from satellite campaigns, reanalysis datasets, and prior modelling studies reduce the forward evaluations needed during calibration.
Scenario: CESM2/CAM6 calibration for marine stratocumulus over the Southeast Pacific. Cloud droplet effective radius constrained using MODIS Collection 6.1 retrievals and in-situ data from the VOCALS-REx campaign (Arica–Iquique transect, 20°S). Aerosol optical depth at 550 nm drawn from AERONET Level 2.0 sites along the Chilean and Peruvian coast.
cloud_droplet_r_eff [µm]
aerosol_optical_depth [–]
02
Pharmacokinetics & Drug Development
Population PK/PD models estimate clearance, absorption rate, and volume of distribution. When patient-level data is limited — pediatric populations, rare diseases, early-phase trials — priors fitted from published studies on related compounds or prior clinical phases provide meaningful constraint.
Scenario: Two-compartment population PK model for vancomycin in neonatal intensive care (gestational age 28–40 weeks, birth weight 0.8–4.5 kg). Clearance and central volume priors informed by published NONMEM analyses in neonatal cohorts (e.g., Kimura et al. 2004; Anderson et al. 2007; Marqués-Miñana et al. 2010). Target: AUC₂₄/MIC-guided dosing for methicillin-resistant Staphylococcus aureus sepsis.
cl [L/h/kg]
vc [L/kg]
03
Hydrology & Water Resources
Conceptual rainfall-runoff models (e.g. SWAT, HBV) have parameters such as hydraulic conductivity and infiltration capacity that vary spatially and are rarely measured directly. Priors can be synthesized from regionalization studies, soil databases, and calibration results for similar catchments.
Scenario: HBV model for the Zala River catchment (Lake Balaton watershed, western Hungary, ~2,600 km²). Dominant soils: Luvisols and Gleysols on Pannonian lacustrine sediments. Field capacity and K_sat priors informed by the EU Hydraulic Properties of European Soils (HYPRES) database and regionalization studies from Central European mesoscale catchments (e.g., Parajka et al. 2007 for Austrian HBV calibrations).
fc [mm]
k_sat [mm/day]
04
Structural & Civil Engineering
Finite element model updating requires estimating stiffness, damping ratios, and degradation rates from sensor measurements. Published material testing data and documented values from comparable structures provide a basis for constructing priors, particularly for structural health monitoring.
Scenario: FE model updating for the Z24 highway bridge (Koppigen–Utzenstorf, Canton Bern, Switzerland) — a post-tensioned concrete box-girder bridge monitored continuously before controlled progressive damage and demolition in 1998. Young's modulus of C40/50 concrete and first-mode damping ratio constrained from the SIMCES benchmark dataset and published Eurocode 2 material test compendia.
youngs_modulus [GPa]
damping_ratio [–]
05
Epidemiology & Disease Modelling
Compartmental models (SIR, SEIR, and variants) require transmission rates, incubation periods, and reporting fractions. During emerging outbreaks, observed data is scarce. Priors fitted from published parameters of related pathogens allow calibration to begin earlier and with greater stability.
Scenario: SEIR model for SARS-CoV-2 (B.1.1.7 / Alpha variant) in England, winter 2020–21. Incubation period (exposure to symptom onset) and infectious period priors drawn from published line-list studies and meta-analyses: Lauer et al. (2020, Annals of Internal Medicine), McAloon et al. (2020), and He et al. (2020, Nature Medicine) for serial interval and viral shedding duration.
incubation_period [days]
infectious_period [days]
06
Astrophysics & Cosmology
Cosmological parameter estimation from CMB power spectra, supernovae, or galaxy surveys involves correlated parameters. Priors from independent observational campaigns help break degeneracies and improve posterior precision when combining datasets.
Scenario: Flat ΛCDM parameter estimation combining Planck 2018 CMB TT,TE,EE+lowE power spectra with Pantheon+ Type Ia supernovae distance moduli. H₀ prior informed by the SH0ES Cepheid–supernova distance ladder (Riess et al. 2022) and the TRGB calibration (Freedman et al. 2021). Ω_m prior informed by BOSS DR12 BAO measurements and DES-Y3 cosmic shear.
H0 [km/s/Mpc]
omega_m [–]
07
Ecology & Population Dynamics
Population models involve growth rates, carrying capacities, and dispersal coefficients that are difficult to estimate from sparse field observations. Meta-analyses and published demographic data from ecologically similar species can be synthesized into prior distributions to stabilize inference.
Scenario: Lotka–Volterra predator–prey model for the Canadian lynx (Lynx canadensis) and snowshoe hare (Lepus americanus) system in the boreal forests of northern Alberta and the Northwest Territories, Canada. Prey intrinsic growth rate and carrying capacity priors drawn from the Hudson's Bay Company fur-return records (1845–1935), Elton & Nicholson (1942), and modern mark-recapture studies by Krebs et al. (2001, Ecosystem Dynamics of the Boreal Forest) from the Kluane Lake region, Yukon.
prey_growth_rate [1/year]
carrying_capacity [individuals/km²]
08
Geophysics & Reservoir Engineering
History matching in reservoir simulation calibrates permeability fields, porosity, and relative permeability curves against production data. The problem is severely underdetermined. Priors informed by well log interpretations, seismic inversions, and core analysis reduce non-uniqueness.
Scenario: Single-phase Darcy flow simulation for the Brent Group (Rannoch–Etive–Ness formations) in the North Sea Brent Province, UK Continental Shelf. Permeability and porosity priors informed by published core-plug analyses from wells 211/29-A and 211/29-B (Ninian field), wireline log interpretations (gamma-ray / density cross-plots), and the Brent Field core data compendium (Morton et al. 1992, Geology of the Brent Group, Geological Society Special Publication No. 61).
permeability [mD]
porosity [–]
09
Robotics & Autonomous Systems
System identification for robot dynamics involves friction coefficients, inertial parameters, and sensor noise characteristics. Published manufacturer data, laboratory benchmarks, and prior deployment studies provide informative priors that improve sim-to-real transfer.
Scenario: Euler–Lagrange rigid-body dynamics identification for the KUKA LBR iiwa 14 R820 7-DOF torque-controlled manipulator. Coulomb friction coefficient and link inertia priors drawn from KUKA's published CAD-derived inertial parameters and from independent identification experiments by Gaz et al. (2019, IEEE RA-L) and Jubien et al. (2014), using exciting sinusoidal joint trajectories with torque-sensor feedback.
coulomb_friction_coeff [–]
link_inertia [kg·m²]
10
Economics & Finance
Bayesian estimation of DSGE models involves structural parameters such as risk aversion, price stickiness (Calvo parameter), and policy reaction coefficients. These are typically informed by microeconometric estimates, and prior specification has a well-documented effect on posterior inference.
Scenario: Three-equation New Keynesian DSGE model (IS curve, Phillips curve, Taylor rule) estimated on US quarterly data (1966:Q1–2019:Q4) from FRED: real GDP growth, GDP deflator inflation, and the effective federal funds rate. Price stickiness duration and inflation response coefficient priors informed by Smets & Wouters (2007, AER), Galí (2015, Monetary Policy, Inflation, and the Business Cycle), and microeconometric price-adjustment frequency estimates from Nakamura & Steinsson (2008, QJE).
price_stickiness_duration [quarters]
inflation_response_coeff [–]