Usage Examples

All referenced scripts in the following section can be found in the /examples directory within repository of the BiasAdjustCXX command-line tool.
The script /examples/example_all_methods.run.sh serves as an example on how to adjust the example data (also provided in the repository) using all implemented methods.
/examples/Hands-On-BiasAdjustCXX.ipynb shows how to clone, build, compile, and run the BiasAdjustCXX command-line tool. Also plots validating the results are presented here.
All methods to bias-adjust climate data can be found in /src/CMethods.cxx. These can be imported into a Jupyter Notebook (with C++ kernel) to test scripts and develop custom algorithms (see /examples/examples.ipynb).
All examples can also be run using the provided Docker image mentioned in section Compilation and Installation.
… More exmples can be found in section: Available Methods

Additive Linear Scaling

Additive Linear Scaling based on means of long-term 31-day intervals applied on a non-stochastic variable like temperatures (“tas”):

Additive Linear Scaling

BiasAdjustCXX                        \
    --ref input_data/observations.nc \ # observations/reference time series of the control period
    --contr input_data/control.nc    \ # simulated time series of the control period
    --scen input_data/scenario.nc    \ # time series to adjust
    -o linear_scaling_result.nc      \ # output file
    -m linear_scaling                \ # adjustment method
    -k "+"                           \ # kind of adjustment ("+" == "add" and "*" == "mult")
    -v tas                           \ # variable to adjust
    -p 4                               # use 4 threads (only if the input data is 3-dimensional)

Note/alternative: The regular linear scaling procedure as described by Teutschbein et al. (2012) needs to be applied on monthly separated data sets. To do so, you have to separate the input data sets into individual long-term months and apply the tool on each of these long-term months using the --no-group flag. This is shown in /examples/example_all_methods.run.sh.

Multiplicative Delta Method

Multiplicative Delta Method based on means of long-term 31-day intervals applied on a stochastic variable like precipitation (“pr”):

Multipliative Delta Method

BiasAdjustCXX                        \
    --ref input_data/observations.nc \
    --contr input_data/control.nc    \
    --scen input_data/scenario.nc    \
    -o delta_method_result.nc        \
    -m delta_method                  \
    -k "*"                           \
    -v pr                            \
    --max-scaling-factor 3             # set custom max-scaling factor to avoid unrealistic results (default: 10)

Note: The multiplicative variant is only preferred when adjusting non-stochastic variables like precipitation, air pressure or wind speed.

Additive Quantile Delta Mapping

Additive Quantile Delta Mapping applied on a non-stochastic variable like temperatures (“tas”):

Additive Quantile Delta Mapping

BiasAdjustCXX                           \
    --ref input_data/observations.nc    \
    --contr input_data/control.nc       \
    --scen input_data/scenario.nc       \
    -o quantile_delta_mapping_result.nc \
    -m quantile_delta_mapping           \
    -k "+"                              \
    -v tas                              \
    -p 4                                \
    -q 250                                # quantiles to respect

Help

Show the help

BiasAdjustCXX --help