# Introduction
This is the readme to a data repository accompanying the publication
Horak J., Hofer M., Maussion F., Gutmann E., Gohm A., Rotach M. W. - **Assessing the Added Value of the Intermediate Complexity Atmospheric Research Model (ICAR)
for Precipitation in Complex Topography**
https://doi.org/10.5194/hess-2018-612
The publication evaluates downscaled precipitation fields (4x4km**2) generated with ICAR with
measurements from several weather stations located along the coast or in the Southern Alps
of the South Island of New Zealand. All methods that yield the data contained in this repository are
described in the publication above.
contact: johannes.horak@uibk.ac.at
## Repository Content
All the data is compressed in one zip file, extraction creates the following folders:
* domain
* froude_number_timeseries
* icar_options
* p24h_timeseries
* scores_mtop4km
* scores_vs_nz
* seasonal_precipitation_maps
* seasonal_precipitation_maps_HR_plots
### folder content quick overview
This provides a brief summary of what is contained in each folder. A more detailed description
is given in the next section.
**domain**
Contains the digital elevation model (`dem_snz_4km.nc`) and a list of all the weather stations
with location data (`stations.csv`). The columns in stations.csv are described in the following:
**froude_number_timeseries**
Contains mean Froude number, atmospheric stability and imaginary/real part of the Brunt-Väisälä
frequncy for each day in the study period that was not excluded due to criteria laid out in
the publication.
**icar_options**
This contains the default options for ICAR (`icar_default.options`) and the ones used for the
simulations that generated the results in this repository (`icar.options`).
**p24h_timeseries**
Here all the 24h accumulated precipitation timeseries (measured, modeled by ICAR and
modeled by ERAI) for each weather station are stored.
**scores_mtop4km**
This folder contains the overall scores (`scores.csv`), seasonal scores (`seasonal_scores.csv`),
weather pattern scores (`pattern_scores.csv`) and scores in dependence of atmospheric stability
and froude number (`froude_scores.csv`).
**scores_vs_nz**
All the results from the model top sensitivity study for the ICAR and ICAR_CP
timeseries are stored here.
**seasonal_precipitation_maps**
Average accumulated 24h precipitation maps for ERAI, ERAI convective precipitation and ICAR
are located here.
**seasonal_precipitation_maps_HR_plots**
High resolution plots (7200x4800) of the seasonal precipitation maps (VCSR, ERAI, ICAR and ICARCP) of
Figure 5 in the publication.
### Detailed description of the folder content
#### folder: domain
Contains the digital elevation model (`dem_snz_4km.nc`) and a list of all the weather stations
with location data (`stations.csv`).
The fields in `dem_snz_4km.nc` are:
`LANDMASK......` landmask indicating ocean (0) or land (1)
`HGT_O.........` unsmoothed topography, elevation in meters
`HGT_M.........` smoothed topography, elevation in meters
`XLONG_M.......` longitude of each grid point
`XLAT_M........` latitude of each grid point
The columns in stations.csv are:
**stations.csv**
`no............` station number
`z.............` station elevation (m)
`station.......` station name
`cluster.......` alpine (A) or coastal (C)
`lat...........` latitude (degrees)
`lon...........` longitude (degrees)
`z_icar........` elevation of ICAR gridcell (m)
`l.............` time series length (years)
`dm............` excluded days due to missing measurements
`Delta.........` distance downwind of main alpine crest (km)
`p1y...........` mean measured annual precipitation (mm)
`p1y_std.......` standard dev. of measured annual precipitation (mm)
`p1y_icar......` mean ICAR_CP modeled annual precipitation (mm)
`p1y_ucar_std..` standard dev. of ICAR_CP modeled annual precipitation (mm)
`p1y_erai......` mean ERAI modeled annual precipitation (mm)
`p1y_erai_std..` standard dev. of ERAI modeled annual precipitation (mm)
`f_cp..........` fraction of convective precipitation in ERAI modeled annual precipitation (mm)
#### folder: froude_number_timeseries
For each day in the 11 year study period that was not excluded due to criteria laid out in
the publication the Brunt Väisälä frequency N and the non-dimensional
mountain height, here referred to as Froude number, was calculated based on atmospheric
conditions in test-volumes to the north-west and south-east of the South Island of New Zealand.
Whether a day is considered as stable or only near stable depends on a threshold for the imaginary
part of N (Nth). Based on this the Froude number is calculated for atmospheric columns that are
considered stable. Therefore, the average Froude number in a testvolume changes when the
threshold is changed.
The threshold is indicated by the filename, e.g. `upstream_fr_1.00e-03.csv` means that
Nth=1.00e-03. Froude numbers were calculated for values of Nth reaching from 0.25e-03 to
6.00e-03 in steps of 0.25e-03.
**upstream_fr_Nth.csv**
`date..........` date (new zealand standard time)
`Fr_mean.......` average Froude number in test-volume (1)
`N_mean........` average Brunt Väisälä frequency in test-volume (s**-1)
`reNm..........` real part of N_mean
`imNm..........` imaginary part of N_mean
#### folder: icar_options
This contains the default options for ICAR (`icar_default.options`) and the ones used for the
simulations that generated the results in this repository (`icar.options`). For more details
comments are provided in the options file or see the official ICAR repository on github
(https://github.com/NCAR/icar).
#### folder: p24h_timeseries
Here all the 24h accumulated precipitation timeseries (measured, modeled by ICAR and
modeled by ERAI) for each weather station are stored. The files contain additional columns
(e.g. weather pattern on that day, icar nearest grid cell, etc.). P24h in the following refers
to the precipitation accumulated during 24 hours and is given in mm. Nearest grid cell refers
to the grid cell clostest to the station location in the grid of the associated model. ERAI is the
abbreviation for the ERA-Interim reanalysis of the ecwmf (https://www.ecmwf.int)
`datetime......` date the 24h accumulated precipitation is associated with
`icar..........` P24h modeled by ICAR at nearest grid cell
`icar_int......` P24h modeled by ICAR interpolated to station coordinates
`erai..........` P24h modeled by ERAI at nearest grid cell
`erai_cp.......` P24h convective precipitation modeled by ERAI at nearest grid cell
`erai_lsp......` P24h large scale precipitation modeled by ERAI at nearest grid cell
`erai_int......` P24h modeled by ERAI interpolated to station coordinates
`erai_cp_int...` P24h convective precipitation by ERAI interpolated to station coordinates
`erai_lsp_int..` P24h large scale precipitation modeled by ERAI interpolated to station coordinates
`meas..........` P24h measured at the weather station
`icar_ref......` P24h modeled by ICAR_CP (=ICAR_INT+ERAI_CP_INT)
`erai_ref......` P24h modeled by ERAI (=ERAI_INT)
#### folder: scores_mtop4km
The files here contain the overall scores (`scores.csv`), seasonal scores (`seasonal_scores.csv`),
weather pattern scores (`pattern_scores.csv`) and scores in dependence of atmospheric stability
and froude number (`froude_scores.csv`).
`scores.csv`, `seasonal_scores.csv` and `pattern_scores.csv` share most of the following
columns:
**score csv files**
`station.......` weather station name
`season........` season for which the quantifiers were calculated (seasonal scores)
`pattern.......` weather pattern for which the quantifiers were calculated (weather pattern scores)
`bootstraps....` number of bootstraps performed
`blocklength...` moving block bootstrap blocklength
`bias_i........` bias of ICAR timeseries compared to measurements
`bias_e........` bias of ERAI timeseries compared to measurements
`ss_mse........` MSE based skill score
`dmse_5perc....` 5th percentile of bootstrapped MSEs
`mse_i.........` MSE of ICAR
`mse_e.........` MSE of ERAI
`hss_1mm.......` HSS for P24h > 1mm
`hss_25mm......` HSS for P24h > 25mm
`hss_50mm......` HSS for P24h > 50mm
`hss_1mm_5perc.` 5th percentile of bootstrapped HSS for P24h > 1mm
`hss_25mm_5perc` 5th percentile of bootstrapped HSS for P24h > 25mm
`hss_50mm_5perc` 5th percentile of bootstrapped HSS for P24h > 50mm
`ss_mse_days...` number of days SS_MSE and bias were calculated for
`hss1mm_days...` number of days HSS for P24h > 1mm was calculated for
`hss25mm_days..` number of days HSS for P24h > 25mm was calculated for
`hss50mm_days..` number of days HSS for P24h > 50mm was calculated for
`a_ix..........` ICAR, number of hits for P24h > x mm with x in {1,25,50}
`b_ix..........` ICAR, number of false hits for P24h > x mm with x in {1,25,50}
`c_ix..........` ICAR, number of missed events for P24h > x mm with x in {1,25,50}
`d_ix..........` ICAR, number of correct misses for P24h > x mm with x in {1,25,50}
`a_ex..........` ERAI, number of hits for P24h > x mm with x in {1,25,50}
`b_ex..........` ERAI, number of false hits for P24h > x mm with x in {1,25,50}
`c_ex..........` ERAI, number of missed events for P24h > x mm with x in {1,25,50}
`d_ex..........` ERAI, number of correct misses for P24h > x mm with x in {1,25,50}
**froude_scores.csv**
`Nth...............` threshold for imaginary part of Brunt Väisälä frequency in 10**-5 s
`total_days........` total number of datapoints (=days) calculations are based on
`mseu_days_rel.....` fraction of near stable days used for corresponding SS_MSE calculation
`msesl_days_rel....` fraction of stable days w. low linearity used for corresponding SS_MSE calculation
`msesg_days_rel....` fraction of stable days w. high linearity used for corresponding SS_MSE calculation
`ss_mseu...........` SS_MSE of near stable days
`ss_msesl..........` SS_MSE of stable days w. low linearity
`ss_msesg..........` SS_MSE of stable days w. high linearity
`hss_u1............` HSS of near stable days where P24h > 1mm
`hss_u25...........` HSS of near stable days where P24h > 25mm
`hss_u50...........` HSS of near stable days where P24h > 50mm
`hss_sl1...........` HSS of stable days w. low linearity where P24h > 1mm
`hss_sl25..........` HSS of stable days w. low linearity where P24h > 25mm
`hss_sl50..........` HSS of stable days w. low linearity where P24h > 50mm
`hss_sg1...........` HSS of stable days w. high linearity where P24h > 1mm
`hss_sg25..........` HSS of stable days w. high linearity where P24h > 25mm
`hss_sg50..........` HSS of stable days w. high linearity where P24h > 50mm
`hss_u1_days_rel...` fraction of near stable days where P24h > 1mm
`hss_u25_days_rel..` fraction of near stable days where P24h > 25mm
`hss_u50_days_rel..` fraction of near stable days where P24h > 50mm
`hss_sl1_days_rel..` fraction of stable days w. low linearity where P24h > 1mm
`hss_sl25_days_rel.` fraction of stable days w. low linearity where P24h > 25mm
`hss_sl50_days_rel.` fraction of stable days w. low linearity where P24h > 50mm
`hss_sg1_days_rel..` fraction of stable days w. high linearity where P24h > 1mm
`hss_sg25_days_rel.` fraction of stable days w. high linearity where P24h > 25mm
`hss_sg50_days_rel.` fraction of stable days w. high linearity where P24h > 50mm
#### folder: scores_vs_nz
The model top sensitivity analysis data is stored in this folder. It contains two subfolders,
`ICAR` and `ICAR_CP`. Both contain the `scores.csv` files (see previous section)
obtained for the reference period 2014-2015 for the different model top settings the simulations
were run for. In the `ICAR_CP` folder ICAR_INT+ERAI_CP_INT timeseries were evaluated while
in the `ICAR` folder ICAR_INT was evaluated. For details see the publication. The columns
contained in the files are essentially the same as described above for `score csv files`.
file list:
* `nz05_scores.csv` (model top height 0.7 km)
* `nz07_scores.csv` (model top height 1.5 km)
* `nz09_scores.csv` (model top height 2.5 km)
* `nz12_scores.csv` (model top height 4.0 km)
* `nz15_scores.csv` (model top height 5.7 km)
* `nz20_scores.csv` (model top height 8.0 km)
#### folder: seasonal_precipitation_maps
The mean precipitation maps, except for the gridded rainfall product (VCSR) by NIWA,
are contained in this folder. All precipitation is given in mm/day.
File list and description:
* `erai_total.nc......` mean P24h of ERAI
* `erai_autumn.nc.....` mean P24h of ERAI during autumn
* `erai_spring.nc.....` mean P24h of ERAI during spring
* `erai_summer.nc.....` mean P24h of ERAI during summer
* `erai_winter.nc.....` mean P24h of ERAI during winter
* `erai_cp_total.nc...` mean P24h convective precipitation of ERAI
* `erai_cp_autumn.nc..` mean P24h convective precipitation of ERAI during autumn
* `erai_cp_spring.nc..` mean P24h convective precipitation of ERAI during spring
* `erai_cp_summer.nc..` mean P24h convective precipitation of ERAI during summer
* `erai_cp_winter.nc..` mean P24h convective precipitation of ERAI during winter
* `icar_total.nc......` mean P24h of ICAR
* `icar_autumn.nc.....` mean P24h of ICAR during autumn
* `icar_spring.nc.....` mean P24h of ICAR during spring
* `icar_summer.nc.....` mean P24h of ICAR during summer
* `icar_winter.nc.....` mean P24h of ICAR during winter
#### folder: seasonal_precipitation_maps_HR_plots
These are high resolution versions of all the panels in Figure 5 of the publication.
File list and description:
* `autumn_erai.png......` difference from the mean P24h of ERAI during autumn
* `autumn_icarcp.png....` difference from the mean P24h of ICARCP during autumn
* `autumn_icar.png......` difference from the mean P24h of ICAR during autumn
* `autumn_vcsr.png......` difference from the mean P24h of VCSR during autumn
* `spring_erai.png......` difference from the mean P24h of ERAI during spring
* `spring_icarcp.png....` difference from the mean P24h of ICARCP during spring
* `spring_icar.png......` difference from the mean P24h of ICAR during spring
* `spring_vcsr.png......` difference from the mean P24h of VCSR during spring
* `summer_erai.png......` difference from the mean P24h of ERAI during summer
* `summer_icarcp.png....` difference from the mean P24h of ICARCP during summer
* `summer_icar.png......` difference from the mean P24h of ICAR during summer
* `summer_vcsr.png......` difference from the mean P24h of VCSR during summer
* `total_erai.png.......` mean P24h of ERAI
* `total_icarcp.png.....` mean P24h of ICARCP
* `total_icar.png.......` mean P24h of ICAR
* `total_vcsr.png.......` mean P24h of VCSR
* `winter_erai.png......` difference from the mean P24h of ERAI during winter
* `winter_icarcp.png....` difference from the mean P24h of ICARCP during winter
* `winter_icar.png......` difference from the mean P24h of ICAR during winter
* `winter_vcsr.png......` difference from the mean P24h of VCSR during winter