Common Workflows

Common Workflows#

This section describes typical workflows for using gwtransport in different scenarios.

Temperature Tracer Test Workflow#

This is the most common workflow for aquifer characterization using temperature data.

Step 1: Data Preparation#

Collect and prepare your data:

import numpy as np
import pandas as pd
from gwtransport.advection import gamma_infiltration_to_extraction

# Load your data
data = pd.read_csv('temperature_data.csv')

# Extract time series
cin_data = data['temp_infiltration'].values
flow_data = data['flow_rate'].values
time_data = data['time'].values

# Create time edges
tedges = np.concatenate([[time_data[0]], time_data])

Step 2: Model Calibration#

Fit model parameters to observed data:

from scipy.optimize import minimize

def objective_function(params):
    mean_vol, std_vol, retardation = params

    # Compute model prediction
    cout_model = gamma_infiltration_to_extraction(
        cin=cin_data,
        flow=flow_data,
        tedges=tedges,
        cout_tedges=tedges,
        mean=mean_vol,
        std=std_vol,
        retardation_factor=retardation,
    )

    # Compare with observations
    error = np.sum((cout_model - cout_observed) ** 2)
    return error

# Initial parameter guess
initial_params = [30000, 8100, 2.0]

# Optimize parameters
result = minimize(objective_function, initial_params, method='Nelder-Mead')
optimal_params = result.x

Step 3: Validation and Prediction#

Validate the model and make predictions:

# Use optimized parameters for predictions
mean_vol, std_vol, retardation = optimal_params

# Predict for new conditions
cout_prediction = gamma_infiltration_to_extraction(
    cin=cin_new,
    flow=flow_new,
    tedges=tedges_new,
    cout_tedges=tedges_new,
    mean=mean_vol,
    std=std_vol,
    retardation_factor=retardation,
)

Streamline Analysis Workflow#

For cases where you have detailed flow field data.

Step 1: Compute Streamlines#

from gwtransport.advection import distribution_infiltration_to_extraction

# Compute areas between streamlines (from flow modeling)
areas_between_streamlines = compute_streamline_areas(flow_field)

# Convert to 3D pore volumes
depth_aquifer = 200.0  # [m]
aquifer_pore_volumes = areas_between_streamlines * depth_aquifer

Step 2: Direct Transport Calculation#

# Use pore volumes directly
cout = distribution_infiltration_to_extraction(
    cin=cin_data,
    flow=flow_data,
    tedges=tedges,
    cout_tedges=tedges,
    aquifer_pore_volumes=aquifer_pore_volumes,
    retardation_factor=1.0,
)

Residence Time Analysis Workflow#

Analyze residence time distributions for different scenarios.

Step 1: Compute Base Residence Times#

from gwtransport.residence_time import compute_residence_time_distribution

# Compute residence time distribution
residence_times = compute_residence_time_distribution(
    aquifer_pore_volumes=aquifer_pore_volumes,
    flow_rates=flow_data,
    tedges=tedges,
)

Step 2: Scenario Analysis#

# Analyze different flow scenarios
scenarios = {
    'low_flow': flow_data * 0.5,
    'normal_flow': flow_data,
    'high_flow': flow_data * 2.0,
}

residence_results = {}
for scenario, flows in scenarios.items():
    residence_results[scenario] = compute_residence_time_distribution(
        aquifer_pore_volumes=aquifer_pore_volumes,
        flow_rates=flows,
        tedges=tedges,
    )

Pathogen Removal Analysis Workflow#

Assess pathogen removal efficiency for treatment design.

Step 1: Define Pathogen Properties#

from gwtransport.logremoval import compute_log_removal

# Define pathogen removal parameters
pathogen_params = {
    'decay_rate': 0.1,  # [1/day]
    'attachment_rate': 0.05,  # [1/day]
    'detachment_rate': 0.01,  # [1/day]
}

Step 2: Compute Removal Efficiency#

# Compute log removal for different residence times
log_removal = compute_log_removal(
    residence_times=residence_times,
    decay_rate=pathogen_params['decay_rate'],
    attachment_rate=pathogen_params['attachment_rate'],
    detachment_rate=pathogen_params['detachment_rate'],
)

# Assess treatment effectiveness
removal_efficiency = 1 - 10**(-log_removal)
print(f"Pathogen removal efficiency: {removal_efficiency:.2%}")

Best Practices#

Data Quality#

  • Ensure high-resolution time series data

  • Check for data gaps and outliers

  • Validate measurement accuracy

  • Consider seasonal variations

Model Validation#

  • Use independent data for validation

  • Check residual patterns

  • Perform sensitivity analysis

  • Compare with physical expectations

Parameter Uncertainty#

  • Quantify parameter uncertainty

  • Propagate uncertainty to predictions

  • Use ensemble approaches when appropriate

  • Document assumptions and limitations

Documentation#

  • Document data sources and processing

  • Record model assumptions

  • Save parameter values and fits

  • Create reproducible workflows

Contents