Operational Seasonal Forecasting Operational Seasonal Forecasting for Bangladesh: for Bangladesh: Application of quantile-to-quantile Application of quantile-to-quantile mapping mapping Tom Hopson Peter Webster Hai-Ru Chang Tom Hopson Peter Webster Hai-Ru Chang Climate Forecast Applications for Climate Forecast Applications for Bangladesh (CFAB) Bangladesh (CFAB)
27
Embed
Operational Seasonal Forecasting for Bangladesh: Application of quantile-to-quantile mapping
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Quantile to Quantile MappingFor 1-, 2-, …, 6-month Precipitation Forecasts
Model Climatology “Observed” Climatology
Pmax
25th 50th 75th 100th
Pfcst
Pre
cipi
tatio
n
Quantile
Pmax
25th 50th 75th 100th
Padj
Quantile
Quantile to Quantile MappingFor 1-, 2-, …, 6-month Discharge Forecasts
Model Precip Climatology “Observed” Q Climatology
Optimal correlation:Brahmaputra discharge 11-day lagged;Ganges discharge: 21 day lagged
The Climate Forecast Applications Project CFAB
Good forecasting skill derived from:1) Spatial scale of the basins2) Satellite-raingauge estimates3) ECMWF forecast skill4) Partnership with FFWC/IWM => Utilize good quality daily border
discharge measurements near-real-time
-- Increase in forecast skill(RMS error) with increasingspatial scale
-- Logarithmic increase
1) Spatial Scale
2) Precipitation Estimates
1) Rain gauge estimates: NOAA CPC and WMO GTS0.5 X 0.5 spatial resolution; 24h temporal resolutionapproximately 100 gauges reporting over combined catchment24hr reporting delay
2) Satellite-derived estimates: Global Precipitation Climatology Project (GPCP)0.25X0.25 spatial resolution; 3hr temporal resolution6hr reporting delaygeostationary infrared “cold cloud top” estimates calibrated from SSM/I and TMI microwave instruments
3) Satellite-derived estimates: NOAA CPC “CMORPH”0.25X0.25 spatial resolution; 3hr temporal resolution18hr reporting delay precipitation rain rates derived from microwave instruments (SSM/I, TMI, AMSU-B), but “cloud tracking” done using infrared satellites
Rain gauge estimates: NOAA CPC and WMO GTS
Comparison of Precipitation Products:
Rain gauge, GPCP, CMORPH, ECMWF
Good comparison for allproducts at large spatial scales
•Hydrology model initial conditions driven by near-real-time GPCP / CMORPH / Raingage precipitation• Ideally, observations would be statistically “just another ensemble member”•Approach: calculate historical NWP-climatology PDF and observation-climatology PDF for each grid using a “kernel” method•For each forecast ensemble, determine its quantile in model-space and extract equivalent quantile in observation-space
ECMWF Ensemble Precipitation Forecast Adjustments -- mapping forecasts from “model-” to “observational-”space
Brahmaputra Catchment-avg Forecasts
Pmax
25th 50th 75th 100th
Pfcst
Pre
cipi
tatio
n
Quantile
Pmax
25th 50th 75th 100th
Padj
Quantile
Quantile to Quantile MappingDone independently for 1-, 2-, …, 10-day forecasts
Model Climatology “Observed” Climatology
Point:Mapping preserves thespatial (and temporal)features of the precipitationforecast fields(i.e. preserves the spatialand temporal covariances)
Original Adjusted
Rank Histogram Comparisons
ECMWF Ensemble Precipitation Forecast Adjustments -- mapping forecasts from “model-” to “observational-”space Brahmaputra Adjusted Forecasts •Benefits:
--Gridded “realistic” forecast values--spatial- and temporal covariances preserved
•Drawbacks:--limited sample set for model-space PDF (2 yrs)--rank histograms show “under-variance”
Mean-Square-Error of the Ensemble-Mean shows skill out to 7-8 days
A Cautionary Warning about using ProbabilisticPrecipitation Forecasts in Hydrologic Modeling
(Importance of Maintaining Spatial and Temporal Covariancesfor Hydrologic Forecasting)
River catchtment A
subB
subC
ensemble1 ensemble2 ensemble3
QBQC
QA
Scenario forsmallest possibleQA? No.
Scenario forlargest possibleQA? No.
QA sameFor all 3 possibleensembles
Scenario foraverage QA?
ConclusionsConclusions
Seasonal Forecasts currently have skill out toSeasonal Forecasts currently have skill out toabout 3 monthsabout 3 months
Possible increased lead-time skill through newPossible increased lead-time skill through newstatistical approachstatistical approach
““Downscaling” (and other methods) holds promise for increased Downscaling” (and other methods) holds promise for increased discharge forecast skilldischarge forecast skill