HYPE Model Guide#

Overview#

HYPE (HYdrological Predictions for the Environment) is a semi-distributed, process-based hydrological model developed by SMHI (Swedish Meteorological and Hydrological Institute). Originally designed for integrated water quality and quantity modeling, HYPE combines robust hydrological process representation with nutrient and contaminant transport capabilities.

Key Capabilities:

  • Semi-distributed structure (subcatchments + soil/land classes)

  • Coupled water and water quality modeling

  • Snow accumulation and melt

  • Soil moisture dynamics

  • Groundwater processes

  • Lake and river routing

  • Nutrient cycling (N, P, C)

  • Optional sediment and contaminant transport

  • Applicable from local to continental scales

Typical Applications:

  • Watershed hydrology

  • Water quality modeling (nutrients, sediments)

  • Climate change impact assessment

  • Land-use change scenarios

  • Flood forecasting

  • Drought analysis

  • Regional to continental-scale assessments

Spatial Scales: Sub-basin (10 km²) to continental (millions of km²)

Temporal Resolution: Daily (standard) or sub-daily

Model Structure#

Spatial Discretization#

HYPE uses a unique two-tier spatial structure:

1. Subcatchments (Sub-basins):

  • Topographically delineated areas

  • Connected through river network

  • Each has elevation, slope, and geographic attributes

2. Soil-Land Classes (SLCs):

  • Within each subcatchment

  • Defined by combinations of: - Soil type (e.g., clay, sand, loam) - Land use (e.g., forest, agriculture, urban) - Elevation zone (optional for snow modeling)

Example Spatial Structure:

Subcatchment 1 (500 km²)
├─ Forest-Sandy_Soil (40%)
├─ Agriculture-Clay_Soil (35%)
├─ Urban-Impervious (15%)
└─ Wetland-Organic_Soil (10%)

Subcatchment 2 (300 km²)
├─ Forest-Loam_Soil (60%)
└─ Agriculture-Clay_Soil (40%)

Process Representation#

Snow and Ice:

  • Degree-day snowmelt

  • Snow distribution by elevation

  • Refreezing and cold content

  • Glacier mass balance (optional)

Soil Water:

  • Up to 3 soil layers

  • Infiltration and percolation

  • Macropore flow

  • Tile drainage

  • Saturation excess and infiltration excess runoff

Evapotranspiration:

  • Temperature-based PET (Priestley-Taylor or Thornthwaite)

  • Actual ET from soil layers

  • Interception evaporation

  • Lake evaporation

Groundwater:

  • Two aquifer levels (shallow, deep)

  • Baseflow generation

  • Aquifer recharge from soil layers

Routing:

  • River routing (kinematic wave or rating curve)

  • Lake retention and routing

  • Delays in subcatchments

  • Flood plains and wetlands

Configuration in SYMFLUENCE#

Model Selection#

HYDROLOGICAL_MODEL: HYPE

Key Configuration Parameters#

Installation and Execution#

Parameter

Default

Description

HYPE_INSTALL_PATH

default

Path to HYPE executable

HYPE_EXE

hype

HYPE executable name

SETTINGS_HYPE_PATH

default

HYPE working directory

SETTINGS_HYPE_INFO

info.txt

Main HYPE configuration file

Calibration Parameters#

Parameter

Default

Description

HYPE_PARAMS_TO_CALIBRATE

null

Comma-separated list of parameters

HYPE_SPINUP_DAYS

365

Spinup period (days)

Input File Specifications#

HYPE uses text-based input files. SYMFLUENCE automatically generates these from your configuration.

Core Configuration Files#

1. info.txt (Main Configuration)

Specifies simulation settings, output variables, and model options:

!!----------------------
!! Simulation settings
!!----------------------
bdate         2010-01-01
cdate         2010-01-01
edate         2020-12-31
tstep         1     ! Daily timestep

!!----------------------
!! Model structure
!!----------------------
substance     0     ! 0 = hydrology only, 1 = include water quality
simN          0     ! Nitrogen simulation (0/1)
simP          0     ! Phosphorus simulation (0/1)
simC          0     ! Carbon simulation (0/1)

!!----------------------
!! Output settings
!!----------------------
outregion     1
outbasin      1
basinoutput   4   ! Daily basin output

!!----------------------
!! Performance criteria
!!----------------------
criterion     1   ! NSE
criterion     2   ! KGE

2. par.txt (Parameters)

Contains model parameters (general + class-specific):

!!----------------------
!! General parameters
!!----------------------
!! parameter     value
gratk            0.100    ! Recession coefficient for runoff (1/day)
rrcs1            0.200    ! Runoff coefficient for soil layer 1
rrcs2            0.100    ! Runoff coefficient for soil layer 2
rrcs3            0.050    ! Runoff coefficient for soil layer 3
ttmp             0.000    ! Threshold temperature for snowmelt (°C)
cmlt             3.500    ! Degree-day factor for snowmelt (mm/°C/day)

!!----------------------
!! Soil class parameters
!!----------------------
!! parameter   soilclass1   soilclass2   soilclass3
wcfc           0.300        0.250        0.200    ! Field capacity
wcwp           0.150        0.120        0.100    ! Wilting point
wcep           0.400        0.350        0.300    ! Effective porosity

3. GeoData.txt (Subcatchment Information)

Attributes for each subcatchment:

SUBID    AREA_km2    ELEV_mean    SLOPE    LAKE_frac    SLC1    SLC1_frac    SLC2    SLC2_frac
1        250.5       450          0.05     0.02         101     0.60         102     0.40
2        180.3       520          0.08     0.00         101     0.45         103     0.55
3        320.1       380          0.03     0.10         102     0.70         104     0.30

Columns:

  • SUBID: Subcatchment ID

  • AREA_km2: Subcatchment area [km²]

  • ELEV_mean: Mean elevation [m]

  • SLOPE: Mean slope [-]

  • LAKE_frac: Fraction covered by lakes

  • SLC#: Soil-land class IDs

  • SLC#_frac: Fraction of subcatchment in each class

4. GeoClass.txt (Soil-Land Class Definitions)

Defines what each SLC represents:

SLCID    SOILTYPE    LANDUSE    CROP    VEGETATION
101      1           1          0       1          ! Sandy soil + Forest
102      2           2          1       2          ! Clay soil + Agriculture + Crop type 1
103      1           3          0       3          ! Sandy soil + Urban
104      3           4          0       4          ! Organic soil + Wetland

5. ForcKey.txt (Forcing Data Mapping)

Maps subcatchments to forcing files:

SUBID    PSFILE    TSFILE    TMAXFILE    TMINFILE
1        P1.txt    T1.txt    TMAX1.txt   TMIN1.txt
2        P2.txt    T2.txt    TMAX2.txt   TMIN2.txt
3        P3.txt    T3.txt    TMAX3.txt   TMIN3.txt

Forcing Data Files#

Precipitation (Pnnn.txt):

DATE            VALUE
2010-01-01      5.2
2010-01-02      0.0
2010-01-03      12.4
...

Temperature (Tnnn.txt):

DATE            VALUE
2010-01-01      2.3
2010-01-02      3.1
2010-01-03      1.8
...

Output File Specifications#

HYPE produces multiple text output files.

Basin Output (basinoutput.txt)#

Time series for each subcatchment:

DATE        SUBID    COUT    ROUT    SNOW    SOIL    EVAP
2010-01-01  1        45.3    42.1    150.2   234.5   0.8
2010-01-01  2        32.1    29.8    180.4   198.2   0.6
2010-01-02  1        42.8    40.2    145.8   232.1   0.9
...

Variables:

  • COUT: Computed outflow [m³/s]

  • ROUT: Routed outflow [m³/s]

  • SNOW: Snow water equivalent [mm]

  • SOIL: Total soil moisture [mm]

  • EVAP: Actual evapotranspiration [mm/day]

Map Output (mapoutput.txt)#

Spatial maps for specific dates:

SUBID    COUT    SNOW    SOIL    SM1    SM2    SM3
1        45.3    150.2   234.5   89.2   98.3   47.0
2        32.1    180.4   198.2   75.1   85.3   37.8
3        52.8    95.3    267.1   102.4  115.2  49.5
...

Time Output (timeoutput.txt)#

Basin-averaged time series:

DATE        PREC    TEMP    EVAP    COUT    SNOW
2010-01-01  5.2     2.3     0.8     45.3    150.2
2010-01-02  0.0     3.1     0.9     42.8    145.8
...

Model-Specific Workflows#

Basic HYPE Workflow#

# config.yaml
DOMAIN_NAME: my_watershed
HYDROLOGICAL_MODEL: HYPE

# Domain definition
DOMAIN_DEFINITION_METHOD: semidistributed
POUR_POINT_COORDS: [15.5, 59.2]  # Sweden

# Spatial discretization
DISCRETIZATION:
  landclass:
    sources: ['MODIS']
  soilclass:
    sources: ['SoilGrids']

# Forcing
FORCING_DATASET: ERA5
FORCING_START_YEAR: 2010
FORCING_END_YEAR: 2020

# HYPE configuration
HYPE_INSTALL_PATH: /path/to/hype
HYPE_SPINUP_DAYS: 365

Run:

symfluence workflow run --config config.yaml

Multi-Land-Use HYPE Application#

For agricultural watersheds:

# config.yaml
HYDROLOGICAL_MODEL: HYPE

# Detailed land and soil discretization
DISCRETIZATION:
  landclass:
    sources: ['MODIS', 'CroplandDataLayer']
    n_classes: 10  # Forest, crops, urban, etc.
  soilclass:
    sources: ['SoilGrids']
    n_classes: 5   # Sand, clay, loam, organic, rocky

# Creates soil-land combinations (up to 50 SLCs)

# Calibrate key parameters
HYPE_PARAMS_TO_CALIBRATE: "gratk,rrcs1,rrcs2,ttmp,cmlt,wcfc,wcwp"

Snow-Focused HYPE#

For snow-dominated regions:

# Add elevation discretization
DISCRETIZATION:
  elevation_bands:
    n_bands: 5
  landclass:
    sources: ['MODIS']
  soilclass:
    sources: ['SoilGrids']

# Snow parameters to calibrate
HYPE_PARAMS_TO_CALIBRATE: "ttmp,cmlt,sfrost,cmrad"

# Parameters:
# ttmp:   Threshold temp for snow/rain (°C)
# cmlt:   Degree-day melt factor (mm/°C/day)
# sfrost: Frost depth parameter
# cmrad:  Radiation melt factor

Large-Scale HYPE (Continental)#

HYPE scales well to large domains:

# E-HYPE style application (European scale)
DOMAIN_DEFINITION_METHOD: merit_basins
MERIT_BASIN_IDS: [1000001, 1000002, ..., 1005000]  # 5000 subcatchments

HYDROLOGICAL_MODEL: HYPE

# Simplified classes for computational efficiency
DISCRETIZATION:
  landclass:
    n_classes: 5  # Major land uses only
  soilclass:
    n_classes: 3  # Coarse, medium, fine

# Parallel calibration
OPTIMIZATION_ALGORITHM: NSGA2
OPTIMIZATION_POPULATION: 100
OPTIMIZATION_MAX_ITERATIONS: 500

Calibration Strategies#

Parameter Bounds#

Typical ranges:

# Runoff and routing
gratk:     [0.01, 0.5]      # Recession coefficient [1/day]
rrcs1:     [0.05, 0.5]      # Runoff coeff layer 1
rrcs2:     [0.01, 0.3]      # Runoff coeff layer 2
rrcs3:     [0.001, 0.1]     # Runoff coeff layer 3

# Soil properties
wcfc:      [0.1, 0.45]      # Field capacity [-]
wcwp:      [0.05, 0.25]     # Wilting point [-]
wcep:      [0.35, 0.55]     # Effective porosity [-]

# Snow parameters
ttmp:      [-2, 2]          # Snow/rain threshold [°C]
cmlt:      [1, 8]           # Melt factor [mm/°C/day]
sfrost:    [5, 50]          # Frost depth [cm]
cmrad:     [0.05, 0.15]     # Radiation factor [-]

# Evapotranspiration
lp:        [0.5, 1.0]       # Limit for PET [-]
cevp:      [0.1, 0.3]       # Correction factor for PET

Calibration Tips#

  1. Spinup is critical:

    HYPE_SPINUP_DAYS: 365  # Minimum 1 year

  2. Soil parameters per class:

    • Can calibrate wcfc, wcwp per soil type

    • Creates many parameters but improves realism

  3. Multi-site calibration:

    • HYPE excels at multi-site (multiple gauges)

    • Helps constrain parameters spatially

  4. Use NSE and KGE together:

    OPTIMIZATION_METRICS: [NSE, KGE, RMSE_log]

  5. Start simple:

    • Begin with 6-8 core parameters

    • Add complexity (class-specific) if needed

Known Limitations#

  1. Complexity:

    • Many input files required

    • Soil-land class setup can be tedious

    • Steep learning curve

  2. Data Requirements:

    • Needs detailed land use and soil maps

    • Lake data required for lake-dominated regions

    • Multiple forcing variables needed

  3. Computational Cost:

    • Daily timestep somewhat slow for large domains

    • Thousands of subcatchments + many SLCs = slow calibration

    • Consider HPC for regional/continental applications

  4. Simplified Processes:

    • Degree-day snowmelt (no energy balance)

    • Simplified groundwater (2 layers)

    • No detailed urban hydrology

  5. Water Quality:

    • Nutrient modules require extensive parameterization

    • Difficult to calibrate hydrology + nutrients simultaneously

Troubleshooting#

Common Issues#

Error: “HYPE executable not found”

# Solution: Verify installation
HYPE_INSTALL_PATH: /absolute/path/to/hype/bin
HYPE_EXE: hype

Error: “GeoData.txt format error”

Check that SYMFLUENCE generated valid GeoData:

# Inspect GeoData.txt
cat <project_dir>/settings/HYPE/GeoData.txt

# Verify:
# - SUBID column is integer
# - AREA_km2 > 0
# - SLC fractions sum to 1.0 per subcatchment

Error: “Missing forcing files”

# Check forcing files exist
ls <project_dir>/forcing/HYPE_input/P*.txt
ls <project_dir>/forcing/HYPE_input/T*.txt

Error: “Simulation crashed - water balance error”

  1. Check parameter values:

    • Ensure wcfc > wcwp

    • Ensure wcep > wcfc

  2. Increase spinup:

    HYPE_SPINUP_DAYS: 730  # 2 years

  3. Check forcing data for gaps or extreme values

Poor performance in snow-dominated regions

# Add elevation discretization
DISCRETIZATION:
  elevation_bands:
    n_bands: 5

# Calibrate snow parameters
HYPE_PARAMS_TO_CALIBRATE: "gratk,rrcs1,rrcs2,ttmp,cmlt,sfrost,cmrad"

Slow execution

Solutions:

  1. Reduce number of SLCs (combine similar classes)

  2. Aggregate small subcatchments

  3. Use coarser time steps if appropriate

  4. Enable parallel HYPE (if compiled with OpenMP/MPI)

Performance Optimization#

Speed up calibration:

  1. Shorter calibration period (5-7 years)

  2. Fewer SLCs (<=10 per subcatchment)

  3. Use DDS or DE algorithms (faster than NSGA2 for HYPE)

Improve accuracy:

  1. Detailed SLC definition (land × soil × elevation)

  2. Multi-site calibration with multiple gauges

  3. Include snow cover data if available

  4. Longer calibration period (10+ years)

Debug HYPE run:

# Run HYPE manually to see detailed output
cd <project_dir>/settings/HYPE/
../../installs/hype/hype

# Check output for warnings/errors

Additional Resources#

HYPE Documentation:

Publications:

SYMFLUENCE-specific:

Example Configurations:

# View HYPE examples
symfluence examples list | grep HYPE

Large-Scale Applications:

Training Materials: