14 th International RSM Workshop September 27 to 29, 2016 Rio de Janeiro, Brazil DOWNSCALING SEASONAL FORECAST SYSTEM USING REGIONAL SPECTRAL MODEL AT FUNCEME Eduardo Svio P. R. Martins Jos Marcelo R. Pereira Francisco das C. Vasconcelos Jnior Ceará Institute for Meteorology and Water Resources
21
Embed
Downscaling seasonal forecast system using regional spectral model at funceme
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.
Transcript
14th International RSM Workshop
September 27 to 29, 2016Rio de Janeiro, Brazil
DOWNSCALING SEASONAL FORECAST SYSTEM USING
REGIONAL SPECTRAL MODEL AT FUNCEME
Eduardo Savio P. R. Martins
Jose Marcelo R. Pereira
Francisco das C. Vasconcelos Junior
Liqiang Sun
Ceará Institute for Meteorology and Water Resources
CONTENTS WHY SEASONAL FORECAST IS IMPORTANT?
CURRENT SITUATION FROM RESERVOIRS
PAST FORECAST SYSTEM
PRESENT FORECAST SYSTEM
RSM97 FORECAST PRODUCTS
FORECAST VERIFICATION
WHY SEASONAL FORECAST IS IMPORTANT?
The Northeast Brazil has faced with high number of severe drought events which have impact on crops planting and agricultural development;
Ceará State is a semiarid region and whose agricultural and socioeconomic developments have been constrained by the shortage of water and the high temporal variability of its supply.
Social, educational and infrastructure programs for droughts relief have also been implemented in the state;
Moreover, improvements on seasonal forecasting methods are still required.
CURRENT SITUATION FROM RESERVOIRS
August 2016
Black 0 – 5%
PAST FORECAST SYSTEM
• FORECAST SYSTEM:• Tests begun in 1999, and• Started operationally in 2001;• Download ECHAM4.5 AGM from IRI;• Climatology 1971-2000; • Dynamic downscaling over Northeast Brazil;• RSM97: ~54 km spatial resolution;• RAMS4.4: ~60 km spatial resolution;• Five-month lead time;
ECHAM4.5PSST AND
SSST
AGM
PRODUCTS
RSM97RAMS4.4
DOWNSCALING
Terciles' probabilities; Probability density
function;Probabilities Map
SMAP;HYDROLOGIC
MODELS
FORECASTS
RESULTSFUNCEMEIRI
PRESENT FORECAST SYSTEM
• FORECAST SYSTEM:• Started operationally in 2012;• ECHAM4.6 AGM running at FUNCEME;• Climatology 1989-2008 (National Multi-model Forecast); • Climatology 1981-2010 (WMO);• Climatology 1971-2000 (RESEARCH);• RSM97: ~54 km spatial resolution;• RSM2008: ~54 km spatial resolution;• RAMS6: ~30 km spatial resolution;• Dynamic Downscaling over Northeast Brazil (RSM97 AND
RAMS6) and South America (RSM2008);• Eight-month lead time;
ECHAM4.620 MEMBERS
PSST (OISSTv2)
AGM
PRODUCTS
RSM97RSM2008*RAMS6*
DOWNSCALING
Terciles' probabilities; Probability density
function;Probabilities Map
SMAP;WASA;
HYDROLOGIC MODELS
FORECASTS
RESULTSFUNCEME
At Development:
RSM2008*RAMS6*
POST-PROCESSING
ECHAM4.6
AGM PRODUCTS
RSM97RSM2008*RAMS6*
DOWNSCALING
• TERCILES' PROBABILITIES; PROBABILITY DENSITY FUNCTION;
RPSS: Measures the improvement of the multi-category probabilistic (below, normal and above) forecast relative to climatology. Perfect score: 1LHSS: Examines the probabilities that were assigned to the later actually observed categories. Perfect score: 1
VERIFICATIONATTRIBUTES DIAGRAM
Reliability is indicated by the proximity of the plotted curve to the diagonal. The deviation from the diagonal gives the conditional bias. If the curve lies below the line, this indicates overforecasting (probabilities too high); points above the line indicate underforecasting (probabilities too low).
VERIFICATIONROC CURVE:
BELOWROC CURVE:
ABOVE
ROC measures the ability of the forecast to discriminate between two alternative outcomes.
Perfect: Curve travels from bottom left to top left of diagram, then across to top right of diagram. Diagonal line indicates no skill.
SUMMARY The forecast verification period 1981-2010 (hindcast) over
Ceará shows high accuracy for forecasting below and above categories events.
Forecast verification issued January valid for the FMAM (2011-2016) season showed that the most likely tercile indicated from forecast was observed in the most of the years.
The interannual variability of rainfall accumulated in Ceará is captured.
The forecast system shows considerable consistency and performance in seasonal forecasting relation to the Brazilian northeast.