Notebook 06: Visualization Hub

This notebook centralizes all visualization for the hybrid bias correction project. It generates diagnostic plots from three domains:

QA Framework (Steps 1–11e): Eight gridded plot types + four station-level breakdowns (region, component×region, province, station)
Taylor Diagrams (Steps 12–19): Product comparison against BMKG stations
Station Validation (Steps 20–23): Spatial maps, multi-threshold curves, and regional box plots from notebook 05 output

A unified batch cell (Step 24) generates all figures across all 36 dekadal periods in a single run.

All plot functions are imported from src/visualisation.py and src/taylor_diagram.py. Each plot type saves into its own sub-folder under figures/.

1 Connect Google Drive (Colab only)

This section is only required when running in Google Colab and your project/data are stored in Google Drive.

Mounting Drive makes your repository and datasets accessible under /content/drive.
If you run this notebook locally (Jupyter / VS Code), skip this section.

Expected structure (Drive) After mounting, your project root should contain: - notebooks/ - src/ - config.yml (or config.yaml)

Proceed to the code cell below to mount Drive.

from google.colab import drive
import os

# Check if the drive is mounted
if os.path.exists("/content/drive"):
    # Try to unmount
    try:
        drive.flush_and_unmount()
        print("Successfully unmounted")
    except:
        print("Unmount failed, the drive might not be mounted or busy")

# Mount the drive
drive.mount("/content/drive")

Successfully unmounted
Mounted at /content/drive

Troubleshooting:
- If Colab becomes disconnected, Reconnect the runtime and rerun the mounting cell. - If we receive an error such as Mountpoint must not already contain files, delete all the sub-folders under “/content/drive” from the Files panel before retrying. We need to delete these one by one starting from the innermost folders, until the last “drive” folder is deleted.

2 Install packages (only if needed)

In most cases, Google Colab already includes the packages required for this workflow. The most common missing dependency is netCDF4 (NetCDF I/O support).

Check what is already installed (Colab)

Before installing anything, you can inspect the current environment by running !pip list.

If all required packages are present and only netCDF4 is missing, install only netCDF4.
If other required packages are missing from !pip list, install them together with netCDF4 in the code cell below.

Local Jupyter note

If you are running in a local environment (Jupyter / VS Code), assume all dependencies were installed when preparing the environment following the main repository README. In that case, you can skip this section.

Proceed to the code cell below only when installation is necessary.

# In Google Colab, almost all packages already available, except netCDF4
!pip install netCDF4

Requirement already satisfied: netCDF4 in /usr/local/lib/python3.12/dist-packages (1.7.4)
Requirement already satisfied: cftime in /usr/local/lib/python3.12/dist-packages (from netCDF4) (1.6.5)
Requirement already satisfied: certifi in /usr/local/lib/python3.12/dist-packages (from netCDF4) (2026.4.22)
Requirement already satisfied: numpy>=1.21.2 in /usr/local/lib/python3.12/dist-packages (from netCDF4) (2.0.2)

3 Quality Assessment Visualization

This section generates spatial maps, distribution plots, and summary charts from the composite quality indices computed by notebook 04. Eight complementary plot types are produced, each highlighting a different aspect of correction quality:

CQI spatial maps – continuous quality index for each correction method
Categorical quality maps – Poor/Fair/Good/Excellent classification
Method improvement map – CQI difference (LSEQM+DL minus LS)
Component quality maps – basic, distribution, temporal sub-scores
Confidence map – assessment reliability
CQI distribution – empirical CDF and histogram
Quality category summary – grouped bar chart of category percentages
Component box plots – score distribution across methods

All plot functions are imported from src/visualisation.py and each type saves into its own sub-folder under figures/qa/.

Step 1: Environment Setup

"""Step 1: Environment setup and configuration."""
import os
import sys
import warnings
import numpy as np
import pandas as pd
import xarray as xr
import matplotlib.pyplot as plt

warnings.filterwarnings("ignore", category=FutureWarning)

# Resolve project root
if os.path.exists("/content/drive/MyDrive/hybrid-bias-correction"):
    project_root = "/content/drive/MyDrive/hybrid-bias-correction"
else:
    project_root = os.path.abspath(os.path.join(os.getcwd(), ".."))

if project_root not in sys.path:
    sys.path.insert(0, project_root)

# ===========================================================================
# AOI config selector
#   'config.yml'      -> full Indonesia (Zenodo input/output bundle)
#   'config_bali.yml' -> Bali example (ships with the repo, ~11 MB)
# Edit this single line to switch the entire pipeline between the two.
# ===========================================================================
CONFIG_FILE = 'config_bali.yml'

from src.config import initialize_config
initialize_config(os.path.join(project_root, CONFIG_FILE))
from src import config

# Output directories for all figure types
figures_dir = os.path.join(config.output_dir, "figures", "qa")
taylor_output_dir = os.path.join(config.output_dir, "figures", "taylor")
station_viz_dir = os.path.join(config.output_dir, "figures", "station_validation")

for d in [figures_dir, taylor_output_dir, station_viz_dir]:
    os.makedirs(d, exist_ok=True)

print(f"Project root:     {project_root}")
print(f"QA figures dir:   {figures_dir}")
print(f"Taylor dir:       {taylor_output_dir}")
print(f"Station viz dir:  {station_viz_dir}")

2026-05-22 04:37:29,235 - INFO - Loaded configuration from /content/drive/MyDrive/hybrid-bias-correction/config_bali.yml
2026-05-22 04:37:29,239 - INFO - Configured main_dir '/content/hybrid-bias-correction' does not exist on this system. Using auto-detected project root: /content/drive/MyDrive/hybrid-bias-correction
2026-05-22 04:37:29,245 - INFO - Configuration initialized successfully.
2026-05-22 04:37:29,245 - INFO -   Main directory: /content/drive/MyDrive/hybrid-bias-correction
2026-05-22 04:37:29,247 - INFO -   Input directory: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali
2026-05-22 04:37:29,248 - INFO -   Output directory: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output
2026-05-22 04:37:29,250 - INFO -   Interactive mode: False
Project root:     /content/drive/MyDrive/hybrid-bias-correction
QA figures dir:   /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa
Taylor dir:       /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor
Station viz dir:  /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/station_validation

Step 2: Select Period and Quality Mode

"""Step 2: User inputs for interactive exploration."""

# Select month (1-12) and dekad (1, 2, or 3)
month = 1
dekad = 1

# Quality mode: "qualitysd" (single-dekad aggregated) or "qualityts" (per-year)
quality_prefix = "qualitysd"

print(f"Period: month {month}, dekad {dekad}")
print(f"Quality mode: {quality_prefix}")

Period: month 1, dekad 1
Quality mode: qualitysd

Step 3: Load Quality Data

"""Step 3: Load QA NetCDF files for all correction methods."""
from src.visualisation import load_quality_data

quality_data = load_quality_data(
    month, dekad,
    quality_prefix=quality_prefix,
    config=config,
)

print(f"Loaded {len(quality_data)} method(s): {list(quality_data.keys())}")

Loaded 3 method(s): ['LS', 'LSEQM', 'LSEQMDL']

Step 4: CQI Spatial Maps

"""Step 4: Continuous Quality Index spatial maps."""
from src.visualisation import plot_cqi_spatial

fig = plot_cqi_spatial(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:40,216 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_cqi_spatial/bali_cli_qa_viz_qualitysd_cqi_spatial_month01_dekad01.png

Step 5: Categorical Quality Maps

"""Step 5: Categorical quality classification maps."""
from src.visualisation import plot_categorical_spatial

fig = plot_categorical_spatial(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:42,770 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_categorical_spatial/bali_cli_qa_viz_qualitysd_categorical_spatial_month01_dekad01.png

Step 6: Method Improvement Map

"""Step 6: CQI improvement map (LSEQM+DL minus LS)."""
from src.visualisation import plot_improvement

fig, imp_stats = plot_improvement(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

Mean improvement:  -0.0369
Median improvement: -0.0374
Pixels improved:   15 (18.8%)
Pixels degraded:   64 (80.0%)
Pixels unchanged:  1 (1.2%)
2026-05-22 04:12:44,800 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_improvement/bali_cli_qa_viz_qualitysd_improvement_month01_dekad01.png

Step 7: Component Quality Maps

"""Step 7: Component quality maps (basic, distribution, temporal)."""
from src.visualisation import plot_components

fig = plot_components(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:47,372 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_components/bali_cli_qa_viz_qualitysd_components_month01_dekad01.png

Step 8: Confidence Map

"""Step 8: Confidence level spatial map."""
from src.visualisation import plot_confidence

fig = plot_confidence(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

Mean confidence:  0.3915
Min confidence:   0.3659
Max confidence:   0.4781
2026-05-22 04:12:49,302 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_confidence/bali_cli_qa_viz_qualitysd_confidence_month01_dekad01.png

Step 9: CQI Distribution Analysis

"""Step 9: Empirical CDF and histogram of CQI."""
from src.visualisation import plot_cqi_distribution

fig = plot_cqi_distribution(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

Median CQI (LS): 0.5947
Median CQI (LSEQM): 0.5531
Median CQI (LSEQM+DL): 0.5536
2026-05-22 04:12:50,815 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_cqi_distribution/bali_cli_qa_viz_qualitysd_cqi_distribution_month01_dekad01.png

Step 10: Quality Category Summary

"""Step 10: Grouped bar chart of quality category percentages."""
from src.visualisation import plot_category_summary

fig, cat_summary = plot_category_summary(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)


Method            Poor      Fair      Good   Excellent       Total
------------------------------------------------------------------
LS                0.0%     56.2%     43.8%        0.0%          80
LSEQM             0.0%     90.0%     10.0%        0.0%          80
LSEQM+DL          0.0%     88.8%     11.2%        0.0%          80
2026-05-22 04:12:51,984 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_category_summary/bali_cli_qa_viz_qualitysd_category_summary_month01_dekad01.png

Step 11: Component Box Plots

"""Step 11: Component score box plots across methods."""
from src.visualisation import plot_component_boxplots

fig = plot_component_boxplots(
    quality_data, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:53,598 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_component_boxplots/bali_cli_qa_viz_qualitysd_component_boxplots_month01_dekad01.png

Step 11b: QA at Station Locations — Regional Summary

Extract gridded QA values at BMKG station locations, then visualize median CQI grouped by the 7 island regions (one bar per method).

"""Step 11b: QA CQI by region — grouped bar chart."""
from src.station_density import load_station_locations
from src.visualisation import plot_qa_regional_bars

station_df = load_station_locations(config.STATION_FILE)

fig = plot_qa_regional_bars(
    quality_data, station_df, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:54,291 - INFO - Loaded 4 station locations from /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/bali_stations_location.csv
2026-05-22 04:12:55,053 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_qa_regional/bali_cli_qa_viz_qualitysd_qa_regional_month01_dekad01.png

Step 11c: QA Components by Region

Box plots of the four QA component scores (CQI, basic statistical, distribution, temporal) at station locations, grouped by island region. Uses the best method (LSEQM+DL) only.

"""Step 11c: QA component box plots by region."""
from src.visualisation import plot_qa_component_by_region

fig = plot_qa_component_by_region(
    quality_data, station_df, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:56,764 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_qa_component_region/bali_cli_qa_viz_qualitysd_qa_component_region_lseqmdl_month01_dekad01.png

Step 11d: QA CQI by Province

Horizontal bar chart of median CQI per province (best method), colour-coded by parent island region.

"""Step 11d: QA CQI by province."""
from src.visualisation import plot_qa_province_bars

fig = plot_qa_province_bars(
    quality_data, station_df, month, dekad,
    quality_prefix=quality_prefix,
    output_dir=figures_dir,
)

2026-05-22 04:12:58,192 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/qa/qualitysd_qa_province/bali_cli_qa_viz_qualitysd_qa_province_lseqmdl_month01_dekad01.png

Step 11e: QA CQI per Station

Per-station CQI bar chart (best method), filtered by region. Change region_filter to view a different island group, or set to None to show all stations.

"""Step 11e: QA CQI per station (single region)."""
from src.visualisation import plot_qa_station_bars

fig = plot_qa_station_bars(
    quality_data, station_df, month, dekad,
    quality_prefix=quality_prefix,
    region_filter="Jawa",  # Change to view other regions, or None for all
    output_dir=figures_dir,
)

4 Taylor Diagram Analysis

This notebook generates Taylor diagrams that visually summarize the performance of bias-corrected precipitation products against independent BMKG station observations.

What is a Taylor Diagram?

A Taylor diagram (Taylor, 2001) simultaneously displays three statistics in a single polar plot:

Pearson correlation (angular axis) – linear agreement with reference
Standard deviation ratio (radial axis) – variability match
Centered RMSE (distance from reference point) – overall error

Each product appears as a marker; perfect agreement places the marker on the reference point (correlation = 1, normalized std = 1). Progressive improvement from raw satellite to fully corrected product is immediately visible as markers move toward the reference.

Products Compared

Product	Description
CPC-UNI	Gauge-based reference (0.50b0, used as correction target)
IMERG-L	Raw satellite precipitation (uncorrected)
IMERG-F	Gauge-adjusted satellite precipitation (uncorrected)
LS	Linear Scaling corrected
LSEQM	LS + EQM with GPD tail adjustment
LSEQM+DL	Full hybrid framework (LS + EQM + GPD + CNN)

Spatial Aggregation Levels

Domain-wide – all 171+ BMKG stations pooled (for the paper)
By island group – 7 regions (Sumatra, Jawa, Kalimantan, etc.)
By province – individual provinces (min 3 stations)
By station – individual station markers with domain aggregate overlay

Prerequisites

Notebook 02 must have completed (corrected files for LS, LSEQM, LSEQMDL)
BMKG station data CSV must be available (location + observations)
CPC-UNI and IMERG-L files must be available

Reference

Taylor, K. E. (2001). Summarizing multiple aspects of model performance in a single diagram. Journal of Geophysical Research, 106(D7), 7183–7192.

Step 12: Compute Taylor Diagram Statistics

This step processes all 36 dekads (12 months × 3 dekads), loading six gridded products (CPC, IMERG-L, IMERG-F, LS, LSEQM, LSEQM+DL) for each dekad and extracting values at each BMKG station location.

Running statistics (sums, sums of squares, cross-products) are accumulated per station per product, enabling exact computation of correlation, standard deviation, and RMSE without keeping all raw data in memory.

Expected runtime: 3–10 minutes depending on disk I/O speed.

"""
Step 12: Compute Taylor statistics across all 36 dekads.

This is the most time-consuming step. The resulting accumulator object
is reused by all subsequent plotting steps.
"""
from src.taylor_diagram import compute_all_taylor_stats

# acc          : pooled per-station accumulator across all 36 dekads
# station_locs : station metadata DataFrame
# acc_by_dekad : {(month, dekad_start): _Accumulator} for per-dekad views
acc, station_locs, acc_by_dekad = compute_all_taylor_stats(config, progress=True)

print(f"\nAccumulator products: {acc.product_keys}")
print(f"Stations: {acc.n_stations}")
print(f"Per-dekad accumulators: {len(acc_by_dekad)}")
print("Total paired observations per product:")
for i, key in enumerate(acc.product_keys):
    total_n = acc.n[i].sum()
    active_stations = (acc.n[i] > 0).sum()
    print(f"  {key:>10s}: {total_n:>12,d} pairs across {active_stations} stations")

  Loaded 4 station locations
  Matched 4/4 stations to obs data
  Station obs: 7670 rows, 2001-01-01 00:00:00 to 2021-12-31 00:00:00, dtype=datetime64[ns]
  Processing dekad  1/36 (month 01, days  1-10) ...    cpc: 250 dates, type=datetime64
    imergl: 250 dates, type=datetime64
    imergf: 250 dates, type=datetime64
    ls: 250 dates, type=datetime64
    lseqm: 250 dates, type=datetime64
    lseqmdl: 250 dates, type=datetime64
  Processing dekad 36/36 (month 12, days 21-31) ...
  Done computing Taylor statistics (184,080 total paired observations).

Accumulator products: ['cpc', 'imergl', 'imergf', 'ls', 'lseqm', 'lseqmdl']
Stations: 4
Per-dekad accumulators: 36
Total paired observations per product:
         cpc:       19,868 pairs across 4 stations
      imergl:       19,871 pairs across 4 stations
      imergf:       19,871 pairs across 4 stations
          ls:       19,871 pairs across 4 stations
       lseqm:       19,871 pairs across 4 stations
     lseqmdl:       19,871 pairs across 4 stations

Step 13: Domain-Wide Taylor Diagram

Pool all station data across the full Indonesian domain. This produces one marker per product, showing the overall performance comparison.

This is the figure intended for the paper (Figure 4 in revision notes).

"""
Step 13: Generate domain-wide Taylor diagram.
"""
from src.taylor_diagram import generate_domain_taylor, print_taylor_stats

fig_domain = generate_domain_taylor(
    acc, station_locs,
    output_dir=taylor_output_dir,
    normalize=True,
    max_std_ratio=2.0,
)
plt.show()


Domain-wide Taylor Statistics (median of per-station)
=====================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.283 |   15.797 |   11.378 |   0.720 |   16.704 |   16.656 |   -1.270 |     19868
IMERG-L      |  0.336 |   15.796 |   11.801 |   0.747 |   16.275 |   16.231 |   -1.208 |     19871
IMERG-F      |  0.332 |   15.796 |   10.713 |   0.678 |   15.953 |   15.875 |   -1.576 |     19871
LS           |  0.326 |   15.796 |   11.922 |   0.755 |   16.429 |   16.398 |   -1.007 |     19871
LSEQM        |  0.326 |   15.796 |   13.555 |   0.858 |   17.148 |   17.137 |   -0.618 |     19871
LSEQM+DL     |  0.328 |   15.796 |   13.450 |   0.852 |   17.068 |   17.058 |   -0.599 |     19871
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide.png

Step 14: Taylor Diagrams by Island Group

Generate a multi-panel figure with one Taylor diagram per major island group: Sumatra, Jawa, Kalimantan, Sulawesi, Bali Nusa Tenggara, Maluku, and Papua. Stations are pooled within each region.

This reveals regional differences in correction performance (e.g., Java with dense gauge coverage vs. Papua with sparse coverage).

"""
Step 14: Generate per-island Taylor diagrams.
"""
from src.taylor_diagram import generate_island_taylor

fig_island = generate_island_taylor(
    acc, station_locs,
    output_dir=taylor_output_dir,
    normalize=True,
)
plt.show()

  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island.png

Step 15: Taylor Diagrams by Province

Generate per-province Taylor diagrams for provinces with at least 3 stations. Provinces with fewer stations are skipped as the pooled statistics would be unreliable.

This provides the finest administrative-level breakdown of correction performance.

"""
Step 15: Generate per-province Taylor diagrams.
"""
from src.taylor_diagram import generate_province_taylor

fig_province = generate_province_taylor(
    acc, station_locs,
    output_dir=taylor_output_dir,
    min_stations=3,
    normalize=True,
)
plt.show()

  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province.png

Step 16: Station-Level Taylor Diagram

Shows individual station markers (small, semi-transparent) for each product, overlaid with larger opaque markers representing the domain-wide aggregate. This reveals the spread of per-station performance within each product category.

Stations that cluster tightly around the aggregate marker indicate consistent correction quality across the network. Wide scatter indicates spatially variable performance.

"""
Step 16: Generate station-level Taylor diagram.
"""
from src.taylor_diagram import generate_station_taylor

fig_station = generate_station_taylor(
    acc, station_locs,
    output_dir=taylor_output_dir,
    normalize=True,
    max_std_ratio=2.0,
)
plt.show()

  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station.png

Step 17: Save Statistics to CSV

Export per-station Taylor statistics for all products to a CSV file. This enables further analysis (e.g., in R, Excel) without re-running the computation.

"""
Step 17: Save per-station Taylor statistics to CSV.
"""
from src.taylor_diagram import save_taylor_stats_csv

csv_path = os.path.join(taylor_output_dir, 'taylor_statistics_per_station.csv')
stats_df = save_taylor_stats_csv(acc, station_locs, csv_path)

print(f"\nCSV shape: {stats_df.shape}")
print(f"Columns: {list(stats_df.columns)}")
stats_df.head(10)

  Saved Taylor statistics CSV: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_statistics_per_station.csv

CSV shape: (4, 48)
Columns: ['station_id', 'station_name', 'lon', 'lat', 'province', 'region', 'cpc_correlation', 'cpc_std_obs', 'cpc_std_prd', 'cpc_rmse', 'cpc_crmse', 'cpc_bias', 'cpc_n', 'imergl_correlation', 'imergl_std_obs', 'imergl_std_prd', 'imergl_rmse', 'imergl_crmse', 'imergl_bias', 'imergl_n', 'imergf_correlation', 'imergf_std_obs', 'imergf_std_prd', 'imergf_rmse', 'imergf_crmse', 'imergf_bias', 'imergf_n', 'ls_correlation', 'ls_std_obs', 'ls_std_prd', 'ls_rmse', 'ls_crmse', 'ls_bias', 'ls_n', 'lseqm_correlation', 'lseqm_std_obs', 'lseqm_std_prd', 'lseqm_rmse', 'lseqm_crmse', 'lseqm_bias', 'lseqm_n', 'lseqmdl_correlation', 'lseqmdl_std_obs', 'lseqmdl_std_prd', 'lseqmdl_rmse', 'lseqmdl_crmse', 'lseqmdl_bias', 'lseqmdl_n']

	station_id	station_name	lon	lat	province	region	cpc_correlation	cpc_std_obs	cpc_std_prd	cpc_rmse	...	lseqm_crmse	lseqm_bias	lseqm_n	lseqmdl_correlation	lseqmdl_std_obs	lseqmdl_std_prd	lseqmdl_rmse	lseqmdl_crmse	lseqmdl_bias	lseqmdl_n
0	97230	Stasiun Meteorologi I Gusti Ngurah Rai	115.17000	-8.75000	Bali	Bali Nusa Tenggara	0.422967	14.076497	11.978168	14.113797	...	15.711541	0.129558	7066	0.388224	14.084034	14.321515	15.712075	15.711541	0.129558	7066
1	97232	Stasiun Geofisika Denpasar	115.21000	-8.67689	Bali	Bali Nusa Tenggara	0.363039	17.173476	13.714254	17.725246	...	18.037177	-0.733262	2839	0.390044	17.173476	15.374826	18.052076	18.037177	-0.733262	2839
2	97234	Pos Pengamatan Kahang-Kahang	115.61083	-8.36500	Bali	Bali Nusa Tenggara	0.202155	14.420211	10.777178	16.198327	...	16.559040	-0.503573	4528	0.268154	14.418665	12.579447	16.403721	16.397118	-0.465405	4528
3	97236	Stasiun Klimatologi Jembrana	114.61760	-8.34100	Bali	Bali Nusa Tenggara	0.176104	17.358224	9.624979	18.551897	...	18.796229	-2.543381	5438	0.202014	17.358199	11.524528	18.967525	18.796229	-2.543381	5438

4 rows × 48 columns

Step 18: One-Click Generation (Alternative)

If you want to generate everything in a single call instead of running Steps 12–17 individually, use generate_all_taylor_diagrams(). This computes statistics and produces all four diagram variants plus the CSV.

Note: Only run this if you have NOT already run Steps 12–17 above, as it will repeat the computation.

"""
Step 18 (Alternative): Generate all Taylor diagrams in one call.

Uncomment and run this cell INSTEAD of Steps 12-17 if preferred.
"""
# from src.taylor_diagram import generate_all_taylor_diagrams
#
# acc, station_locs, acc_by_dekad = generate_all_taylor_diagrams(
#     config,
#     output_dir=taylor_output_dir,
# )

Step 19: Custom Taylor Diagram (On-Demand)

Generate a Taylor diagram for a specific subset of products or a specific region. Modify the parameters below as needed.

Examples: - Show only corrected products (exclude raw IMERG) - Focus on a single island - Compare only LSEQM vs LSEQM+DL

"""
Step 19: Custom Taylor diagram for a specific region or product subset.

Modify TARGET_REGION and PRODUCTS_TO_SHOW as needed.
"""
from src.taylor_diagram import plot_taylor_diagram, print_taylor_stats

# --- User settings ---
TARGET_REGION = 'Jawa'  # One of: Sumatra, Jawa, Kalimantan, Sulawesi,
                        #         Bali Nusa Tenggara, Maluku, Papua
                        # Set to None for domain-wide

PRODUCTS_TO_SHOW = None  # None = all products, or e.g. ['ls', 'lseqm', 'lseqmdl']

# --- Compute stats ---
if TARGET_REGION is not None:
    mask = (station_locs['Region'] == TARGET_REGION).values
    n_st = mask.sum()
    stats = acc.compute_median(station_mask=mask)
    title = f'Taylor Diagram: {TARGET_REGION} ({n_st} stations)'
else:
    stats = acc.compute_median()
    title = 'Taylor Diagram: Domain-wide'

# --- Print table ---
print_taylor_stats(stats, title=title)

# --- Plot ---
fig, ax = plot_taylor_diagram(
    stats,
    title=title,
    normalize=True,
    products_to_show=PRODUCTS_TO_SHOW,
    max_std_ratio=2.0,
)
plt.show()

5 Station Validation Visualization

This section visualizes the independent station validation results from notebook 05. Three plot types are generated from the CSV outputs:

Station metric scatter maps – spatial distribution of 6 key metrics
WMO multi-threshold curves – categorical skill across intensity thresholds
Regional box plots – metric distributions by main island

All plot functions are imported from src/visualisation.py and save into sub-folders under figures/station_validation/.

Prerequisites: Notebook 05 batch must have completed (CSV outputs in data/output/station_validation/).

Step 20: Load Station Validation CSV

Load per-station validation metrics and multi-threshold summaries from the CSV files produced by notebook 05. Set month and dekad to match the period selected in Step 2.

"""Step 20: Load station validation CSV data for a single period."""
from src.station_density import load_station_locations
from src.station_validation import merge_station_metadata

# Load station locations
station_df = load_station_locations(config.STATION_FILE)

# Build path to station validation CSV
sv_dir = getattr(config, "STATION_VALIDATION_OUTPUT_DIR", None)
if sv_dir is None:
    sv_dir = os.path.join(config.output_dir, "station_validation")

month_str = f"{month:02d}"
dekad_map = {1: "01", 2: "11", 3: "21"}
dekad_str = dekad_map[dekad]

# Load metrics for each method
sv_metrics = {}
for method_key, method_abbr in [("LS", "ls"), ("LSEQM", "lseqm"), ("LSEQMDL", "lseqmdl")]:
    csv_path = os.path.join(
        sv_dir,
        f"station_validation_{method_abbr}_month{month_str}_dekad{dekad_str}.csv",
    )
    if os.path.exists(csv_path):
        sv_metrics[method_key] = pd.read_csv(csv_path, index_col=0)
        print(f"  {method_key}: {len(sv_metrics[method_key])} stations from {csv_path}")
    else:
        print(f"  {method_key}: not found -- {csv_path}")

# Load multi-threshold summaries
sv_mt_summaries = {}
for method_key, method_abbr in [("LS", "ls"), ("LSEQM", "lseqm"), ("LSEQMDL", "lseqmdl")]:
    mt_csv = os.path.join(
        sv_dir,
        f"multi_threshold_summary_{method_abbr}_month{month_str}_dekad{dekad_str}.csv",
    )
    if os.path.exists(mt_csv):
        sv_mt_summaries[method_key] = pd.read_csv(mt_csv, index_col=0)

print(f"\nMetrics loaded: {list(sv_metrics.keys())}")
print(f"Multi-threshold summaries: {list(sv_mt_summaries.keys())}")

2026-05-22 04:39:33,654 - INFO - Loaded 4 station locations from /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/bali_stations_location.csv
  LS: 4 stations from /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/station_validation/station_validation_ls_month01_dekad01.csv
  LSEQM: 4 stations from /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/station_validation/station_validation_lseqm_month01_dekad01.csv
  LSEQMDL: 4 stations from /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/station_validation/station_validation_lseqmdl_month01_dekad01.csv

Metrics loaded: ['LS', 'LSEQM', 'LSEQMDL']
Multi-threshold summaries: ['LS', 'LSEQM', 'LSEQMDL']

Step 21: Station Metric Scatter Maps

Spatial scatter plots showing the distribution of 6 key validation metrics (correlation, NSE, relative bias, RMSE, CSI, POD) at each BMKG station location for the best correction method.

"""Step 21: Station metric scatter maps."""
from src.visualisation import plot_station_metric_maps

# Use best method available
sv_best = "LSEQMDL" if "LSEQMDL" in sv_metrics else list(sv_metrics.keys())[-1]

if sv_metrics:
    fig = plot_station_metric_maps(
        sv_metrics[sv_best], station_df, month, dekad,
        method_name=sv_best,
        output_dir=station_viz_dir,
    )
else:
    print("No station validation data available.")

2026-05-22 04:39:53,007 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/station_validation/station_validation_metric_maps/bali_cli_qa_viz_station_validation_metric_maps_lseqmdl_month01_dekad01.png

Step 22: Multi-Threshold Performance Curves

WMO/TD-No. 1485 style performance curves showing how categorical verification scores (POD, FAR, CSI, FBI, ETS, HSS) change across 7 precipitation intensity thresholds (1–150 mm/day). Each line represents a correction method.

"""Step 22: WMO multi-threshold performance curves."""
from src.visualisation import plot_multi_threshold_curves

if sv_mt_summaries:
    fig = plot_multi_threshold_curves(
        sv_mt_summaries, month, dekad,
        output_dir=station_viz_dir,
    )
else:
    print("No multi-threshold summaries available.")

2026-05-22 04:40:04,516 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/station_validation/station_validation_threshold_curves/bali_cli_qa_viz_station_validation_threshold_curves_month01_dekad01.png

2026-05-22 04:40:06,863 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/station_validation/station_validation_exceedance_frequency/bali_cli_qa_viz_station_validation_exceedance_frequency_month01_dekad01.png

Step 23: Regional Box Plots

Box plots of key validation metrics grouped by main island (Region). Reveals regional differences in correction quality — e.g., densely gauged Java vs. sparsely gauged Papua.

"""Step 23: Regional box plots of station validation metrics."""
from src.visualisation import plot_regional_boxplots
from src.station_validation import merge_station_metadata

if sv_metrics:
    sv_best = "LSEQMDL" if "LSEQMDL" in sv_metrics else list(sv_metrics.keys())[-1]
    regional_df = merge_station_metadata(sv_metrics[sv_best], station_df)
    if "Region" in regional_df.columns:
        fig = plot_regional_boxplots(
            regional_df, month, dekad,
            method_name=sv_best, group_col="Region",
            output_dir=station_viz_dir,
        )
    else:
        print("Region column not found in station metadata.")
else:
    print("No station validation data available.")

2026-05-22 04:40:13,487 - INFO - Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/station_validation/station_validation_regional_region/bali_cli_qa_viz_station_validation_regional_region_lseqmdl_month01_dekad01.png

Summary

This notebook generated diagnostic visualizations from three complementary evaluation domains:

QA Framework (Steps 4–11e): 8 gridded plot types + 4 station-level regional/province/station breakdowns from composite quality indices
Taylor Diagrams (Steps 12–19): Product comparison against BMKG stations
Station Validation (Steps 20–23): Spatial maps, threshold curves, and regional box plots from independent validation

All figures are saved into organized sub-folders under figures/. The unified batch below generates all figures for all 36 periods.

Step 24: Unified Batch – All Visualization

Generate all figures across all 36 dekadal periods in a single run. This calls four batch orchestrators:

run_qa_batch_viz() — 8 QA plot types × 36 periods × 2 quality modes
run_qa_regional_batch_viz() — QA regional/province/station plots × 36 periods × 2 modes
generate_all_taylor_diagrams() — 4 Taylor diagram variants + CSV
run_station_validation_batch_viz() — 3 station validation plot types × 36 periods

Prerequisites: Only Step 1 (setup) is required. Steps 2–23 can be skipped for batch-only runs.

"""Step 24: Unified batch -- all visualization for all 36 periods."""
from src.visualisation import (
    run_qa_batch_viz,
    run_qa_regional_batch_viz,
    run_station_validation_batch_viz,
)
from src.taylor_diagram import generate_all_taylor_diagrams

# # --- 1. QA Framework plots (8 types x 36 periods x 2 modes) ---
# print("=" * 60)
# print("BATCH 1/4: QA Framework Visualization")
# print("=" * 60)
# qa_summaries = {}
# for _qp in ["qualitysd", "qualityts"]:
#     print(f"\n--- Quality mode: {_qp} ---")
#     qa_summaries[_qp] = run_qa_batch_viz(
#         quality_prefix=_qp,
#         output_dir=figures_dir,
#         config=config,
#         progress=True,
#     )

# # --- 2. QA Regional / Province / Station plots (x 2 modes) ---
# print("\n" + "=" * 60)
# print("BATCH 2/4: QA Regional Breakdown")
# print("=" * 60)
# qa_regional_summaries = {}
# for _qp in ["qualitysd", "qualityts"]:
#     print(f"\n--- Quality mode: {_qp} ---")
#     qa_regional_summaries[_qp] = run_qa_regional_batch_viz(
#         quality_prefix=_qp,
#         output_dir=figures_dir,
#         config=config,
#         progress=True,
#     )

# --- 3. Taylor Diagrams ---
print("\n" + "=" * 60)
print("BATCH 3/4: Taylor Diagram Generation")
print("=" * 60)
try:
    td_acc, td_locs = generate_all_taylor_diagrams(
        config,
        output_dir=taylor_output_dir,
    )
    print("Taylor diagrams generated successfully.")
except Exception as e:
    print(f"Taylor diagram generation failed: {e}")

# # --- 4. Station Validation plots (3 types x 36 periods) ---
# print("\n" + "=" * 60)
# print("BATCH 4/4: Station Validation Visualization")
# print("=" * 60)
# sv_summary = run_station_validation_batch_viz(
#     output_dir=station_viz_dir,
#     config=config,
#     progress=True,
# )

print("\n" + "=" * 60)
print("ALL BATCHES COMPLETE")
print("=" * 60)


============================================================
BATCH 3/4: Taylor Diagram Generation
============================================================
============================================================
  Taylor Diagram Generation
============================================================

Step 1/8: Computing statistics across all dekads ...
  Loaded 4 station locations
  Matched 4/4 stations to obs data
  Station obs: 7670 rows, 2001-01-01 00:00:00 to 2021-12-31 00:00:00, dtype=datetime64[ns]
  Processing dekad  1/36 (month 01, days  1-10) ...    cpc: 250 dates, type=datetime64
    imergl: 250 dates, type=datetime64
    imergf: 250 dates, type=datetime64
    ls: 250 dates, type=datetime64
    lseqm: 250 dates, type=datetime64
    lseqmdl: 250 dates, type=datetime64
  Processing dekad 36/36 (month 12, days 21-31) ...
  Done computing Taylor statistics (184,080 total paired observations).

Step 2/8: Saving pooled statistics CSV ...
  Saved Taylor statistics CSV: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_statistics_per_station.csv

Step 3/8: Saving per-dekad statistics CSV ...
  Saved per-dekad Taylor statistics CSV: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_statistics_per_dekad.csv

Step 4/8: Domain-wide Taylor diagram (pooled) ...

Domain-wide Taylor Statistics (median of per-station)
=====================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.283 |   15.797 |   11.378 |   0.720 |   16.704 |   16.656 |   -1.270 |     19868
IMERG-L      |  0.336 |   15.796 |   11.801 |   0.747 |   16.275 |   16.231 |   -1.208 |     19871
IMERG-F      |  0.332 |   15.796 |   10.713 |   0.678 |   15.953 |   15.875 |   -1.576 |     19871
LS           |  0.326 |   15.796 |   11.922 |   0.755 |   16.429 |   16.398 |   -1.007 |     19871
LSEQM        |  0.326 |   15.796 |   13.555 |   0.858 |   17.148 |   17.137 |   -0.618 |     19871
LSEQM+DL     |  0.328 |   15.796 |   13.450 |   0.852 |   17.068 |   17.058 |   -0.599 |     19871
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide.png

Step 5/8: By-island Taylor diagrams (pooled) ...
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island.png

Step 6/8: By-province Taylor diagrams (pooled) ...
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province.png

Step 7/8: Per-dekad Taylor diagrams ...

  Per-dekad summary (median across 36 dekads) ...

Domain-wide Taylor Statistics — Per-Dekad (median of per-station)
=================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.307 |   12.959 |    8.845 |   0.683 |   13.310 |   13.261 |   -1.132 |     19868
IMERG-L      |  0.282 |   12.959 |   10.610 |   0.819 |   14.272 |   14.242 |   -0.924 |     19871
IMERG-F      |  0.279 |   12.959 |    9.266 |   0.715 |   13.724 |   13.670 |   -1.221 |     19871
LS           |  0.282 |   12.959 |    9.611 |   0.742 |   13.808 |   13.782 |   -0.851 |     19871
LSEQM        |  0.265 |   12.959 |   11.517 |   0.889 |   14.897 |   14.884 |   -0.619 |     19871
LSEQM+DL     |  0.271 |   12.959 |   11.429 |   0.882 |   14.789 |   14.776 |   -0.623 |     19871
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_per_dekad.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_per_dekad.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_per_dekad.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_per_dekad.png

  Per-dekad individual diagrams (36 periods) ...
    Dekad  1/36 (month01_dekad01) ...
Domain-wide Taylor Statistics — Month01 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.365 |   20.335 |   17.106 |   0.841 |   21.293 |   21.265 |   -1.090 |       578
IMERG-L      |  0.399 |   20.335 |   16.320 |   0.803 |   20.418 |   20.365 |   -1.477 |       578
IMERG-F      |  0.365 |   20.335 |   15.405 |   0.758 |   20.703 |   20.553 |   -2.488 |       578
LS           |  0.399 |   20.335 |   17.647 |   0.868 |   20.943 |   20.934 |   -0.622 |       578
LSEQM        |  0.401 |   20.335 |   18.126 |   0.891 |   21.127 |   21.126 |    0.220 |       578
LSEQM+DL     |  0.401 |   20.335 |   18.126 |   0.891 |   21.124 |   21.123 |    0.230 |       578
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month01_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month01_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month01_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month01_dekad01.png
    Dekad  2/36 (month01_dekad11) ...
Domain-wide Taylor Statistics — Month01 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.225 |   21.562 |   16.221 |   0.752 |   23.956 |   23.890 |   -1.775 |       584
IMERG-L      |  0.309 |   21.562 |   12.217 |   0.567 |   21.475 |   21.241 |   -3.159 |       584
IMERG-F      |  0.289 |   21.562 |   12.373 |   0.574 |   21.818 |   21.538 |   -3.483 |       584
LS           |  0.309 |   21.562 |   14.811 |   0.687 |   22.108 |   22.061 |   -1.445 |       584
LSEQM        |  0.298 |   21.562 |   17.184 |   0.797 |   23.260 |   23.226 |   -1.265 |       584
LSEQM+DL     |  0.298 |   21.562 |   16.998 |   0.788 |   23.167 |   23.132 |   -1.265 |       584
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month01_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month01_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month01_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month01_dekad11.png
    Dekad  3/36 (month01_dekad21) ...
Domain-wide Taylor Statistics — Month01 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.179 |   19.821 |   15.173 |   0.765 |   22.789 |   22.708 |   -1.928 |       662
IMERG-L      |  0.271 |   19.821 |   14.571 |   0.735 |   21.312 |   21.185 |   -2.320 |       662
IMERG-F      |  0.265 |   19.821 |   14.593 |   0.736 |   21.383 |   21.277 |   -2.128 |       662
LS           |  0.271 |   19.821 |   16.452 |   0.830 |   22.108 |   22.067 |   -1.336 |       662
LSEQM        |  0.280 |   19.821 |   17.409 |   0.878 |   22.439 |   22.423 |   -0.836 |       662
LSEQM+DL     |  0.279 |   19.821 |   17.409 |   0.878 |   22.448 |   22.432 |   -0.836 |       662
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month01_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month01_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month01_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month01_dekad21.png
    Dekad  4/36 (month02_dekad01) ...
Domain-wide Taylor Statistics — Month02 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.178 |   20.905 |   14.656 |   0.701 |   23.480 |   23.292 |   -2.964 |       610
IMERG-L      |  0.225 |   20.905 |   18.719 |   0.895 |   24.746 |   24.720 |   -1.134 |       610
IMERG-F      |  0.210 |   20.905 |   17.216 |   0.824 |   24.195 |   24.129 |   -1.779 |       610
LS           |  0.225 |   20.905 |   16.471 |   0.788 |   23.662 |   23.518 |   -2.600 |       610
LSEQM        |  0.212 |   20.905 |   16.840 |   0.806 |   23.995 |   23.901 |   -2.119 |       610
LSEQM+DL     |  0.212 |   20.905 |   16.840 |   0.806 |   23.995 |   23.901 |   -2.119 |       610
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month02_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month02_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month02_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month02_dekad01.png
    Dekad  5/36 (month02_dekad11) ...
Domain-wide Taylor Statistics — Month02 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.156 |   16.231 |   14.539 |   0.896 |   20.139 |   20.024 |   -2.146 |       593
IMERG-L      |  0.189 |   16.231 |   12.659 |   0.780 |   18.635 |   18.604 |   -1.070 |       593
IMERG-F      |  0.185 |   16.231 |   12.448 |   0.767 |   18.586 |   18.538 |   -1.346 |       593
LS           |  0.189 |   16.231 |   13.405 |   0.826 |   19.042 |   18.999 |   -1.275 |       593
LSEQM        |  0.186 |   16.231 |   17.206 |   1.060 |   21.349 |   21.345 |   -0.452 |       593
LSEQM+DL     |  0.186 |   16.231 |   17.111 |   1.054 |   21.285 |   21.282 |   -0.383 |       593
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month02_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month02_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month02_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month02_dekad11.png
    Dekad  6/36 (month02_dekad21) ...
Domain-wide Taylor Statistics — Month02 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.179 |   19.589 |   14.259 |   0.728 |   22.125 |   22.068 |   -1.588 |       490
IMERG-L      |  0.175 |   19.553 |   14.186 |   0.726 |   22.275 |   22.053 |   -3.137 |       493
IMERG-F      |  0.196 |   19.553 |   13.103 |   0.670 |   21.635 |   21.298 |   -3.804 |       493
LS           |  0.175 |   19.553 |   16.154 |   0.826 |   23.166 |   23.077 |   -2.021 |       493
LSEQM        |  0.167 |   19.553 |   17.450 |   0.892 |   23.963 |   23.941 |   -1.043 |       493
LSEQM+DL     |  0.162 |   19.553 |   17.450 |   0.892 |   24.033 |   24.010 |   -1.043 |       493
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month02_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month02_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month02_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month02_dekad21.png
    Dekad  7/36 (month03_dekad01) ...
Domain-wide Taylor Statistics — Month03 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.071 |   18.094 |   14.169 |   0.783 |   22.232 |   22.172 |   -1.633 |       560
IMERG-L      |  0.196 |   18.094 |   13.826 |   0.764 |   20.544 |   20.512 |   -1.159 |       560
IMERG-F      |  0.182 |   18.094 |   13.681 |   0.756 |   20.608 |   20.601 |   -0.524 |       560
LS           |  0.196 |   18.094 |   13.640 |   0.754 |   20.479 |   20.419 |   -1.570 |       560
LSEQM        |  0.188 |   18.094 |   15.869 |   0.877 |   21.716 |   21.706 |   -0.651 |       560
LSEQM+DL     |  0.188 |   18.094 |   15.869 |   0.877 |   21.716 |   21.706 |   -0.651 |       560
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month03_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month03_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month03_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month03_dekad01.png
    Dekad  8/36 (month03_dekad11) ...
Domain-wide Taylor Statistics — Month03 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.317 |   16.726 |   10.369 |   0.620 |   17.098 |   16.652 |   -3.880 |       560
IMERG-L      |  0.317 |   16.726 |   13.833 |   0.827 |   18.121 |   18.014 |   -1.967 |       560
IMERG-F      |  0.309 |   16.726 |   12.288 |   0.735 |   17.503 |   17.426 |   -1.640 |       560
LS           |  0.317 |   16.726 |   14.590 |   0.872 |   18.539 |   18.385 |   -2.382 |       560
LSEQM        |  0.380 |   16.726 |   13.625 |   0.815 |   17.173 |   17.097 |   -1.611 |       560
LSEQM+DL     |  0.380 |   16.726 |   13.625 |   0.815 |   17.167 |   17.091 |   -1.611 |       560
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month03_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month03_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month03_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month03_dekad11.png
    Dekad  9/36 (month03_dekad21) ...
Domain-wide Taylor Statistics — Month03 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.313 |   13.806 |   10.618 |   0.769 |   14.563 |   14.549 |   -0.640 |       627
IMERG-L      |  0.334 |   13.806 |   11.562 |   0.837 |   14.757 |   14.750 |   -0.451 |       627
IMERG-F      |  0.292 |   13.806 |   11.513 |   0.834 |   15.178 |   15.178 |    0.045 |       627
LS           |  0.334 |   13.806 |   11.958 |   0.866 |   14.952 |   14.940 |   -0.597 |       627
LSEQM        |  0.333 |   13.806 |   12.095 |   0.876 |   15.026 |   15.025 |   -0.174 |       627
LSEQM+DL     |  0.327 |   13.806 |   11.939 |   0.865 |   15.012 |   15.011 |   -0.174 |       627
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month03_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month03_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month03_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month03_dekad21.png
    Dekad 10/36 (month04_dekad01) ...
Domain-wide Taylor Statistics — Month04 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.212 |   13.080 |    9.689 |   0.741 |   14.628 |   14.532 |   -1.678 |       542
IMERG-L      |  0.268 |   13.080 |    9.695 |   0.741 |   14.110 |   14.042 |   -1.384 |       542
IMERG-F      |  0.257 |   13.080 |    9.970 |   0.762 |   14.299 |   14.266 |   -0.972 |       542
LS           |  0.268 |   13.080 |   10.747 |   0.822 |   14.556 |   14.536 |   -0.758 |       542
LSEQM        |  0.246 |   13.080 |   12.602 |   0.963 |   15.774 |   15.771 |   -0.293 |       542
LSEQM+DL     |  0.246 |   13.080 |   12.602 |   0.963 |   15.772 |   15.771 |   -0.197 |       542
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month04_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month04_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month04_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month04_dekad01.png
    Dekad 11/36 (month04_dekad11) ...
Domain-wide Taylor Statistics — Month04 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.336 |   13.043 |    9.869 |   0.757 |   13.532 |   13.450 |   -1.484 |       546
IMERG-L      |  0.309 |   13.043 |   10.691 |   0.820 |   14.121 |   14.077 |   -1.110 |       546
IMERG-F      |  0.299 |   13.043 |    9.080 |   0.696 |   13.602 |   13.484 |   -1.782 |       546
LS           |  0.309 |   13.043 |   10.070 |   0.772 |   13.870 |   13.794 |   -1.450 |       546
LSEQM        |  0.338 |   13.043 |   10.954 |   0.840 |   13.934 |   13.914 |   -0.745 |       546
LSEQM+DL     |  0.337 |   13.043 |   11.031 |   0.846 |   13.974 |   13.954 |   -0.745 |       546
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month04_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month04_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month04_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month04_dekad11.png
    Dekad 12/36 (month04_dekad21) ...
Domain-wide Taylor Statistics — Month04 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.139 |    8.238 |    6.216 |   0.755 |    9.622 |    9.603 |   -0.608 |       531
IMERG-L      |  0.219 |    8.238 |    7.363 |   0.894 |    9.773 |    9.773 |   -0.097 |       531
IMERG-F      |  0.209 |    8.238 |    4.912 |   0.596 |    8.696 |    8.665 |   -0.734 |       531
LS           |  0.219 |    8.238 |    5.626 |   0.683 |    8.940 |    8.900 |   -0.850 |       531
LSEQM        |  0.230 |    8.238 |    6.201 |   0.753 |    9.139 |    9.102 |   -0.817 |       531
LSEQM+DL     |  0.230 |    8.238 |    6.201 |   0.753 |    9.139 |    9.102 |   -0.817 |       531
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month04_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month04_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month04_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month04_dekad21.png
    Dekad 13/36 (month05_dekad01) ...
Domain-wide Taylor Statistics — Month05 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.406 |   10.517 |    6.617 |   0.629 |    9.897 |    9.893 |   -0.273 |       524
IMERG-L      |  0.352 |   10.517 |   10.880 |   1.034 |   12.191 |   12.187 |    0.286 |       524
IMERG-F      |  0.335 |   10.517 |   10.250 |   0.975 |   11.986 |   11.975 |    0.524 |       524
LS           |  0.352 |   10.517 |    6.845 |   0.651 |   10.372 |   10.337 |   -0.852 |       524
LSEQM        |  0.350 |   10.517 |    6.852 |   0.652 |   10.383 |   10.352 |   -0.803 |       524
LSEQM+DL     |  0.350 |   10.517 |    6.852 |   0.652 |   10.381 |   10.352 |   -0.788 |       524
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month05_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month05_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month05_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month05_dekad01.png
    Dekad 14/36 (month05_dekad11) ...
Domain-wide Taylor Statistics — Month05 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.376 |    8.604 |    4.616 |   0.537 |    8.126 |    8.093 |   -0.730 |       515
IMERG-L      |  0.466 |    8.604 |   10.008 |   1.163 |    9.719 |    9.690 |    0.756 |       515
IMERG-F      |  0.468 |    8.604 |    7.588 |   0.882 |    8.402 |    8.394 |    0.378 |       515
LS           |  0.466 |    8.604 |    6.447 |   0.749 |    7.994 |    7.992 |   -0.198 |       515
LSEQM        |  0.408 |    8.604 |    6.550 |   0.761 |    8.423 |    8.422 |    0.050 |       515
LSEQM+DL     |  0.409 |    8.604 |    6.278 |   0.730 |    8.325 |    8.325 |   -0.001 |       515
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month05_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month05_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month05_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month05_dekad11.png
    Dekad 15/36 (month05_dekad21) ...
Domain-wide Taylor Statistics — Month05 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.433 |   11.588 |    8.565 |   0.739 |   11.057 |   11.035 |   -0.695 |       566
IMERG-L      |  0.284 |   11.588 |    8.382 |   0.723 |   12.234 |   12.220 |   -0.594 |       566
IMERG-F      |  0.266 |   11.588 |    9.452 |   0.816 |   12.891 |   12.861 |   -0.879 |       566
LS           |  0.284 |   11.588 |    8.243 |   0.711 |   12.163 |   12.163 |   -0.097 |       566
LSEQM        |  0.254 |   11.588 |   12.378 |   1.068 |   14.664 |   14.654 |    0.528 |       566
LSEQM+DL     |  0.266 |   11.588 |   12.212 |   1.054 |   14.439 |   14.428 |    0.565 |       566
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month05_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month05_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month05_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month05_dekad21.png
    Dekad 16/36 (month06_dekad01) ...
Domain-wide Taylor Statistics — Month06 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.023 |   10.851 |    3.143 |   0.290 |   11.324 |   11.228 |   -1.472 |       513
IMERG-L      | -0.012 |   10.851 |    2.973 |   0.274 |   11.403 |   11.284 |   -1.641 |       513
IMERG-F      | -0.018 |   10.851 |    3.587 |   0.331 |   11.581 |   11.490 |   -1.451 |       513
LS           | -0.012 |   10.851 |    2.821 |   0.260 |   11.363 |   11.243 |   -1.648 |       513
LSEQM        | -0.004 |   10.851 |    4.681 |   0.431 |   11.927 |   11.834 |   -1.489 |       513
LSEQM+DL     | -0.003 |   10.851 |    4.681 |   0.431 |   11.925 |   11.832 |   -1.485 |       513
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month06_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month06_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month06_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month06_dekad01.png
    Dekad 17/36 (month06_dekad11) ...
Domain-wide Taylor Statistics — Month06 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.400 |    8.218 |    6.995 |   0.851 |    8.396 |    8.393 |    0.222 |       513
IMERG-L      |  0.205 |    8.218 |    6.897 |   0.839 |    9.586 |    9.586 |    0.041 |       513
IMERG-F      |  0.183 |    8.218 |    5.647 |   0.687 |    9.081 |    9.081 |   -0.048 |       513
LS           |  0.205 |    8.218 |    5.869 |   0.714 |    9.069 |    9.068 |   -0.149 |       513
LSEQM        |  0.168 |    8.218 |    6.750 |   0.821 |    9.724 |    9.719 |    0.331 |       513
LSEQM+DL     |  0.171 |    8.218 |    6.750 |   0.821 |    9.710 |    9.704 |    0.333 |       513
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month06_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month06_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month06_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month06_dekad11.png
    Dekad 18/36 (month06_dekad21) ...
Domain-wide Taylor Statistics — Month06 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.342 |   11.715 |    8.473 |   0.723 |   11.903 |   11.882 |   -0.709 |       510
IMERG-L      |  0.281 |   11.715 |   12.998 |   1.110 |   14.935 |   14.856 |   -1.537 |       510
IMERG-F      |  0.302 |   11.715 |   11.347 |   0.969 |   13.699 |   13.626 |   -1.420 |       510
LS           |  0.281 |   11.715 |   14.032 |   1.198 |   15.560 |   15.551 |   -0.511 |       510
LSEQM        |  0.343 |   11.715 |   12.055 |   1.029 |   13.643 |   13.631 |   -0.568 |       510
LSEQM+DL     |  0.345 |   11.715 |   11.826 |   1.010 |   13.481 |   13.470 |   -0.530 |       510
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month06_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month06_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month06_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month06_dekad21.png
    Dekad 19/36 (month07_dekad01) ...
Domain-wide Taylor Statistics — Month07 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.381 |    4.544 |    3.560 |   0.783 |    4.582 |    4.581 |    0.071 |       519
IMERG-L      |  0.475 |    4.544 |    3.743 |   0.824 |    4.300 |    4.300 |   -0.077 |       519
IMERG-F      |  0.498 |    4.544 |    4.045 |   0.890 |    4.327 |    4.323 |    0.170 |       519
LS           |  0.475 |    4.544 |    5.699 |   1.254 |    5.362 |    5.339 |    0.491 |       519
LSEQM        |  0.449 |    4.544 |    7.952 |   1.750 |    7.205 |    7.170 |    0.714 |       519
LSEQM+DL     |  0.448 |    4.544 |    7.616 |   1.676 |    6.933 |    6.901 |    0.669 |       519
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month07_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month07_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month07_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month07_dekad01.png
    Dekad 20/36 (month07_dekad11) ...
Domain-wide Taylor Statistics — Month07 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.313 |   10.426 |    5.153 |   0.494 |   10.174 |   10.083 |   -1.358 |       523
IMERG-L      |  0.267 |   10.426 |    4.183 |   0.401 |   10.324 |   10.144 |   -1.916 |       523
IMERG-F      |  0.285 |   10.426 |    4.160 |   0.399 |   10.208 |   10.065 |   -1.707 |       523
LS           |  0.267 |   10.426 |    4.674 |   0.448 |   10.327 |   10.224 |   -1.461 |       523
LSEQM        |  0.293 |   10.426 |    4.652 |   0.446 |   10.199 |   10.096 |   -1.445 |       523
LSEQM+DL     |  0.293 |   10.426 |    4.652 |   0.446 |   10.199 |   10.096 |   -1.445 |       523
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month07_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month07_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month07_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month07_dekad11.png
    Dekad 21/36 (month07_dekad21) ...
Domain-wide Taylor Statistics — Month07 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.544 |    5.454 |    4.928 |   0.904 |    4.994 |    4.977 |   -0.414 |       563
IMERG-L      |  0.537 |    5.454 |    5.859 |   1.074 |    5.469 |    5.457 |   -0.368 |       563
IMERG-F      |  0.532 |    5.454 |    4.759 |   0.873 |    4.992 |    4.978 |   -0.370 |       563
LS           |  0.537 |    5.454 |    8.714 |   1.598 |    7.399 |    7.394 |    0.252 |       563
LSEQM        |  0.544 |    5.454 |    6.063 |   1.112 |    5.525 |    5.523 |   -0.158 |       563
LSEQM+DL     |  0.544 |    5.454 |    6.063 |   1.112 |    5.525 |    5.523 |   -0.158 |       563
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month07_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month07_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month07_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month07_dekad21.png
    Dekad 22/36 (month08_dekad01) ...
Domain-wide Taylor Statistics — Month08 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.301 |    5.565 |    3.595 |   0.646 |    5.661 |    5.643 |   -0.444 |       519
IMERG-L      |  0.314 |    5.565 |    2.098 |   0.377 |    5.366 |    5.296 |   -0.865 |       519
IMERG-F      |  0.350 |    5.565 |    2.191 |   0.394 |    5.281 |    5.220 |   -0.799 |       519
LS           |  0.314 |    5.565 |    3.595 |   0.646 |    5.630 |    5.599 |   -0.595 |       519
LSEQM        |  0.350 |    5.565 |    4.383 |   0.788 |    5.785 |    5.755 |   -0.587 |       519
LSEQM+DL     |  0.354 |    5.565 |    4.383 |   0.788 |    5.766 |    5.735 |   -0.596 |       519
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month08_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month08_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month08_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month08_dekad01.png
    Dekad 23/36 (month08_dekad11) ...
Domain-wide Taylor Statistics — Month08 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.352 |    3.888 |    3.237 |   0.833 |    4.101 |    4.092 |   -0.276 |       528
IMERG-L      |  0.166 |    3.888 |    1.437 |   0.370 |    3.962 |    3.915 |   -0.608 |       528
IMERG-F      |  0.198 |    3.888 |    1.772 |   0.456 |    3.969 |    3.941 |   -0.465 |       528
LS           |  0.166 |    3.888 |    1.658 |   0.426 |    3.980 |    3.966 |   -0.342 |       528
LSEQM        |  0.152 |    3.888 |    3.724 |   0.958 |    4.966 |    4.960 |   -0.253 |       528
LSEQM+DL     |  0.152 |    3.888 |    3.724 |   0.958 |    4.965 |    4.957 |   -0.271 |       528
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month08_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month08_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month08_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month08_dekad11.png
    Dekad 24/36 (month08_dekad21) ...
Domain-wide Taylor Statistics — Month08 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.672 |    6.283 |    4.252 |   0.677 |    4.653 |    4.653 |   -0.003 |       559
IMERG-L      |  0.242 |    6.283 |    1.501 |   0.239 |    6.142 |    6.096 |   -0.751 |       559
IMERG-F      |  0.266 |    6.283 |    1.711 |   0.272 |    6.105 |    6.056 |   -0.770 |       559
LS           |  0.242 |    6.283 |    2.988 |   0.476 |    6.280 |    6.269 |   -0.367 |       559
LSEQM        |  0.251 |    6.283 |    4.009 |   0.638 |    6.565 |    6.551 |   -0.427 |       559
LSEQM+DL     |  0.253 |    6.283 |    3.854 |   0.613 |    6.501 |    6.487 |   -0.427 |       559
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month08_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month08_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month08_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month08_dekad21.png
    Dekad 25/36 (month09_dekad01) ...
Domain-wide Taylor Statistics — Month09 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.521 |    4.562 |    3.015 |   0.661 |    4.017 |    3.945 |   -0.760 |       509
IMERG-L      |  0.684 |    4.562 |    5.243 |   1.149 |    3.966 |    3.945 |    0.403 |       509
IMERG-F      |  0.706 |    4.562 |    3.447 |   0.756 |    3.269 |    3.239 |   -0.441 |       509
LS           |  0.684 |    4.562 |    3.808 |   0.835 |    3.447 |    3.397 |   -0.584 |       509
LSEQM        |  0.709 |    4.562 |    4.566 |   1.001 |    3.492 |    3.483 |   -0.262 |       509
LSEQM+DL     |  0.709 |    4.562 |    4.474 |   0.981 |    3.458 |    3.449 |   -0.262 |       509
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month09_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month09_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month09_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month09_dekad01.png
    Dekad 26/36 (month09_dekad11) ...
Domain-wide Taylor Statistics — Month09 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.255 |   14.398 |   10.522 |   0.731 |   15.536 |   15.515 |    0.801 |       508
IMERG-L      |  0.355 |   14.398 |   12.746 |   0.885 |   15.483 |   15.480 |   -0.342 |       508
IMERG-F      |  0.330 |   14.398 |    8.703 |   0.604 |   14.201 |   14.158 |   -1.095 |       508
LS           |  0.355 |   14.398 |   16.862 |   1.171 |   17.939 |   17.874 |    1.528 |       508
LSEQM        |  0.437 |   14.398 |   18.305 |   1.271 |   17.878 |   17.667 |    2.739 |       508
LSEQM+DL     |  0.447 |   14.398 |   18.167 |   1.262 |   17.649 |   17.416 |    2.854 |       508
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month09_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month09_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month09_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month09_dekad11.png
    Dekad 27/36 (month09_dekad21) ...
Domain-wide Taylor Statistics — Month09 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.312 |   12.876 |    8.300 |   0.645 |   12.959 |   12.959 |    0.039 |       497
IMERG-L      |  0.329 |   12.876 |    9.077 |   0.705 |   13.085 |   13.084 |    0.108 |       497
IMERG-F      |  0.380 |   12.876 |    6.177 |   0.480 |   11.982 |   11.981 |   -0.147 |       497
LS           |  0.329 |   12.876 |    9.198 |   0.714 |   13.129 |   13.129 |   -0.049 |       497
LSEQM        |  0.325 |   12.876 |   13.116 |   1.019 |   15.141 |   15.100 |    1.110 |       497
LSEQM+DL     |  0.325 |   12.876 |   13.036 |   1.012 |   15.096 |   15.054 |    1.137 |       497
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month09_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month09_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month09_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month09_dekad21.png
    Dekad 28/36 (month10_dekad01) ...
Domain-wide Taylor Statistics — Month10 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.251 |    8.303 |    4.646 |   0.560 |    8.554 |    8.438 |   -1.403 |       524
IMERG-L      |  0.147 |    8.303 |    6.739 |   0.812 |    9.922 |    9.896 |   -0.706 |       524
IMERG-F      |  0.170 |    8.303 |    7.377 |   0.889 |   10.133 |   10.125 |   -0.425 |       524
LS           |  0.147 |    8.303 |    7.866 |   0.947 |   10.568 |   10.567 |   -0.174 |       524
LSEQM        |  0.170 |    8.303 |    9.769 |   1.177 |   11.703 |   11.698 |    0.351 |       524
LSEQM+DL     |  0.170 |    8.303 |    9.482 |   1.142 |   11.500 |   11.494 |    0.349 |       524
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month10_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month10_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month10_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month10_dekad01.png
    Dekad 29/36 (month10_dekad11) ...
Domain-wide Taylor Statistics — Month10 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.282 |   10.998 |    6.580 |   0.598 |   11.206 |   11.108 |   -1.480 |       508
IMERG-L      |  0.287 |   10.998 |    7.158 |   0.651 |   11.501 |   11.270 |   -2.294 |       508
IMERG-F      |  0.275 |   10.998 |    5.601 |   0.509 |   11.100 |   10.885 |   -2.174 |       508
LS           |  0.287 |   10.998 |    8.265 |   0.751 |   11.870 |   11.708 |   -1.952 |       508
LSEQM        |  0.254 |   10.998 |    8.036 |   0.731 |   12.011 |   11.861 |   -1.890 |       508
LSEQM+DL     |  0.252 |   10.998 |    8.015 |   0.729 |   12.011 |   11.862 |   -1.890 |       508
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month10_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month10_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month10_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month10_dekad11.png
    Dekad 30/36 (month10_dekad21) ...
Domain-wide Taylor Statistics — Month10 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.313 |    9.715 |    8.841 |   0.910 |   10.898 |   10.895 |   -0.231 |       570
IMERG-L      |  0.285 |    9.715 |   10.529 |   1.084 |   12.125 |   12.123 |   -0.226 |       570
IMERG-F      |  0.283 |    9.715 |   10.004 |   1.030 |   11.820 |   11.811 |   -0.452 |       570
LS           |  0.285 |    9.715 |    9.157 |   0.943 |   11.294 |   11.294 |    0.109 |       570
LSEQM        |  0.245 |    9.715 |   10.979 |   1.130 |   12.774 |   12.755 |    0.706 |       570
LSEQM+DL     |  0.245 |    9.715 |   10.925 |   1.125 |   12.742 |   12.719 |    0.773 |       570
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month10_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month10_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month10_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month10_dekad21.png
    Dekad 31/36 (month11_dekad01) ...
Domain-wide Taylor Statistics — Month11 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.438 |   16.060 |    8.848 |   0.551 |   14.784 |   14.550 |   -2.623 |       539
IMERG-L      |  0.351 |   16.060 |   11.563 |   0.720 |   16.494 |   16.160 |   -3.302 |       539
IMERG-F      |  0.374 |   16.060 |    7.563 |   0.471 |   15.419 |   14.972 |   -3.684 |       539
LS           |  0.351 |   16.060 |   10.511 |   0.654 |   16.057 |   15.805 |   -2.835 |       539
LSEQM        |  0.277 |   16.060 |   10.269 |   0.639 |   16.717 |   16.491 |   -2.740 |       539
LSEQM+DL     |  0.279 |   16.060 |   10.269 |   0.639 |   16.689 |   16.476 |   -2.660 |       539
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month11_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month11_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month11_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month11_dekad01.png
    Dekad 32/36 (month11_dekad11) ...
Domain-wide Taylor Statistics — Month11 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.150 |   15.418 |    9.088 |   0.589 |   16.711 |   16.685 |   -0.944 |       554
IMERG-L      |  0.107 |   15.418 |    9.123 |   0.592 |   17.149 |   17.056 |   -1.780 |       554
IMERG-F      |  0.120 |   15.418 |    7.963 |   0.516 |   16.644 |   16.479 |   -2.336 |       554
LS           |  0.107 |   15.418 |    6.852 |   0.444 |   16.328 |   16.190 |   -2.119 |       554
LSEQM        |  0.107 |   15.418 |    8.802 |   0.571 |   17.035 |   16.912 |   -2.038 |       554
LSEQM+DL     |  0.109 |   15.418 |    8.690 |   0.564 |   16.978 |   16.855 |   -2.038 |       554
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month11_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month11_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month11_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month11_dekad11.png
    Dekad 33/36 (month11_dekad21) ...
Domain-wide Taylor Statistics — Month11 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.093 |   18.100 |   12.824 |   0.709 |   21.220 |   21.187 |   -1.174 |       586
IMERG-L      |  0.225 |   18.100 |   12.140 |   0.671 |   19.417 |   19.394 |   -0.945 |       586
IMERG-F      |  0.241 |   18.100 |   12.378 |   0.684 |   19.406 |   19.308 |   -1.953 |       586
LS           |  0.225 |   18.100 |   10.024 |   0.554 |   18.780 |   18.614 |   -2.491 |       586
LSEQM        |  0.207 |   18.100 |   15.782 |   0.872 |   21.452 |   21.406 |   -1.408 |       586
LSEQM+DL     |  0.207 |   18.100 |   15.782 |   0.872 |   21.453 |   21.407 |   -1.408 |       586
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month11_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month11_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month11_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month11_dekad21.png
    Dekad 34/36 (month12_dekad01) ...
Domain-wide Taylor Statistics — Month12 Dekad01 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.112 |   20.725 |   14.793 |   0.714 |   24.146 |   24.082 |   -1.750 |       582
IMERG-L      |  0.132 |   20.725 |   14.321 |   0.691 |   23.597 |   23.584 |   -0.775 |       582
IMERG-F      |  0.120 |   20.725 |   13.500 |   0.651 |   23.452 |   23.338 |   -2.307 |       582
LS           |  0.132 |   20.725 |   14.549 |   0.702 |   23.760 |   23.697 |   -1.721 |       582
LSEQM        |  0.149 |   20.725 |   17.182 |   0.829 |   24.896 |   24.873 |   -1.078 |       582
LSEQM+DL     |  0.149 |   20.725 |   16.881 |   0.814 |   24.728 |   24.703 |   -1.112 |       582
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month12_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month12_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month12_dekad01.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month12_dekad01.png
    Dekad 35/36 (month12_dekad11) ...
Domain-wide Taylor Statistics — Month12 Dekad11 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.129 |   19.990 |   15.279 |   0.764 |   23.643 |   23.547 |   -2.128 |       613
IMERG-L      |  0.241 |   19.990 |   16.121 |   0.806 |   22.477 |   22.459 |   -0.903 |       613
IMERG-F      |  0.267 |   19.990 |   12.522 |   0.626 |   20.752 |   20.561 |   -2.805 |       613
LS           |  0.241 |   19.990 |   15.639 |   0.782 |   22.292 |   22.219 |   -1.797 |       613
LSEQM        |  0.232 |   19.990 |   17.055 |   0.853 |   23.112 |   23.068 |   -1.423 |       613
LSEQM+DL     |  0.232 |   19.990 |   16.997 |   0.850 |   23.086 |   23.045 |   -1.380 |       613
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month12_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month12_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month12_dekad11.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month12_dekad11.png
    Dekad 36/36 (month12_dekad21) ...
Domain-wide Taylor Statistics — Month12 Dekad21 (median of per-station)
=======================================================================
Product      |   Corr |    σ_obs |    σ_prd | σ ratio |     RMSE |    CRMSE |     Bias |         N
--------------------------------------------------------------------------------------------------
CPC-UNI      |  0.217 |   18.391 |   21.147 |   1.150 |   24.864 |   24.831 |   -1.284 |       643
IMERG-L      |  0.300 |   18.391 |   14.544 |   0.791 |   19.823 |   19.736 |   -1.853 |       643
IMERG-F      |  0.288 |   18.391 |   11.884 |   0.646 |   19.145 |   18.801 |   -3.615 |       643
LS           |  0.300 |   18.391 |   14.323 |   0.779 |   19.757 |   19.637 |   -2.175 |       643
LSEQM        |  0.276 |   18.391 |   20.057 |   1.091 |   23.190 |   23.172 |   -0.909 |       643
LSEQM+DL     |  0.276 |   18.391 |   19.888 |   1.081 |   23.081 |   23.063 |   -0.909 |       643
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_domain_wide_month12_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_island_month12_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_province_month12_dekad21.png
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station_month12_dekad21.png


Step 8/8: Station-level Taylor diagram (pooled) ...
  Saved: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor/taylor_by_station.png

All Taylor diagrams saved to: /content/drive/MyDrive/hybrid-bias-correction/data/example_bali/output/figures/taylor
Taylor diagram generation failed: too many values to unpack (expected 2)

============================================================
ALL BATCHES COMPLETE
============================================================

Step 25: Batch Summary

"""Step 25: Print aggregated statistics from all batch runs."""
from src.visualisation import print_batch_summary, print_station_validation_viz_summary

print("=== QA Framework Batch ===")
for _qp, _summary in qa_summaries.items():
    print(f"\n--- {_qp} ---")
    print_batch_summary(_summary)

print("\n=== Station Validation Batch ===")
print_station_validation_viz_summary(sv_summary)