DSSAT crop yield evaluation

Crop Evaluation Outcome

A lightweight project page for reviewing the county-level maize yield evaluation products generated by s2s crop-evaluate. The page references the PNG/PDF outputs directly, including the method-specific evaluation folders for the three detrending methods and supported coefficient-selection modes.

A
Experiment ABlue background bars
Select both experiments, then draw.
B
Experiment BRed foreground bars
Model: CWRF Mode: Standard Change: absolute Method: multiplicative Trending Ref: sliding-trend Optfil: fixed reference year Space: target Period: 2012-2025 Issue month: May Folder: ../../models/CWRF/crop/evaluation

Monthly Crop Forecast Skill Summary

Open PNG Open PDF Open CSV
Evaluation data CSV not found at the selected path. After running s2s crop-evaluate, set the evaluation folder to the directory containing the period barplot CSV (for example crop_eval_monthly_barplot_2012-2025_target-multiplicative-refyear-sliding-trend.csv) and the month_MM folders.
Cross-month county median and interquartile skill summary for anomaly correlation, RMSE, normalized RMSE, bias, and Kling-Gupta efficiency.

State Mean Yield Time Series

Open BC CSV Open Raw CSV
Observed (NASS)
Bias Corrected
Raw
Time series of simulated state mean yields (solid colored lines) versus NASS observed state mean yields (solid black lines).

County Metric Maps: May

All maps are generated by s2s crop-evaluate; this page only displays the existing files.

Forcing Diagnostics

Debug-experiment comparison of model CWRFdssatClim_* forcing against reference forcing.

Comparisons Bias Corrected and Raw occupy separate centers; Experiment A is blue behind Experiment B, which is red.
Multi-Model The default Multimodel ensemble averages all available initialization dates and experiments within each model first, then averages across models. Rolling-window or custom member-set ensembles are written to their own named multimodel folders.
Maps Per-month panels include Raw and Bias Corrected maps for ACC, RMSE, nRMSE, and Bias, their differences, and comparative box plots.
Difference The difference figure is calculated as Bias Corrected minus Raw for each county metric.
Method Matrix Use the method, coefficient-mode, and evaluation-space controls to browse multiplicative, linear, and TA10 outputs for fixed-reference, previous-year, or causal fixed-window coefficients.
Default model outputs are expected under <workspace>/models/<MODEL>/crop/evaluation/, and this page builds paths automatically from the selected model and mode.