Use of Intra Seasonal and Seasonal Forecasts to Reduce Risk in - - PowerPoint PPT Presentation

Agricultural and Biological Engineering Agricultural and Biological Engineering

Use of Intra‐Seasonal and Seasonal Forecasts to Reduce Risk in Regional Public Water Supply Management

Chris Martinez University of Florida

Agricultural and Biological Engineering

Overview

• Project Partners
• Project Background & Goals
• Methods
• Results
• Lessons Learned
• Relevance

Funded by NOAA’s Climate Program Office SARP-Water program

Agricultural and Biological Engineering

Background

• Current sources:
• Groundwater (13 wellfields)
• Tampa Bypass Canal/Hillsborough River
• Alafia River
• C.W. Bill Young off‐stream reservoir
• Desalination Plant

1998 2008 2012

Percentage of Water by Source

100% 61% 11% 28% 45.5% 45.5% 9%

Groundwater Permit

192 MGD 158 MGD 121 MGD 90 MGD Pre‐1998 1998 2002 2008

(12-month moving average)

Agricultural and Biological Engineering

Source Allocation Decisions

• Multiple decision scales:
• Water year plan (6‐months prior)
• Month to month adjustments
• Operational decisions (weekly)
• Multiple constraints:
• Permitted groundwater (12‐month moving average)
• Minimum streamflows
• Streamflow extraction ratio (maintain Fluoride limit)
• Costs of different sources

Agricultural and Biological Engineering

Project Goals

• Integrate forecast information into decision making

– Multiple temporal scales – Relevant spatial scales – Integrate forecasts into suite of models used by Tampa Bay Water

• System‐wide Decision Support

– What is the system‐wide benefit/risk of adopting forecast information? – What is the reliability of the current system? – Value judgments under different scenarios?

Agricultural and Biological Engineering

Links Between Climate/Hydrologic Information and Decisions by Tampa Bay Water

Decision Required Information Climate/Hydrology

Set Prices and monthly Source Allocation for water-year Update water-year Allocations

Time-Scale

18 months in advance Monthly,

• ut to 12

months Operational Allocations Weekly,

• ut to 4

weeks

• Estimate of initial reservoir volume
• Scenarios of historical conditions
• Demand forecasts
• Precipitation and Streamflow

forecasts

• Precipitation, Streamflow and

Demand forecasts

Agricultural and Biological Engineering

“Typical” Water Year

• Estimate of end of year reservoir level needed for plan for next water year

Oct --- Nov --- Dec --- Jan --- Feb --- Mar --- Apr --- May --- Jun --- Jul --- Aug --- Sep

Reservoir Max Using Reservoir Filling Reservoir Reservoir Min Max Groundwater use Max Direct Surface Water use

• Greater than expected groundwater pumping impacts next water year plan
• Seasonal forecasts can be used to determine expected

higher/lower groundwater pumping in winter months

Agricultural and Biological Engineering

Different Products for Different Time‐ Scales

• Operational – Ensemble Precipitation,

Streamflow and Demand forecasts derived from medium‐range forecast products

• Monthly/Seasonal – Probabilistic Precipitation

and Streamflow/Withdrawal Climate‐based Forecasts

• Water‐Year – Decision Support, taking into

account previous and next 12 months

Agricultural and Biological Engineering

Operational Time‐scale

Current Forecast Historical Forecast Analogs Historical Observations (Stations or NARR)

• Forecast analogs using

the ESRL/PSD GFS Retrospective forecast archive

• 1‐15 day
• 2.5° x 2.5°
• 1979‐present
• Analog selection can be

tailored to need

Ensemble of Hydrologic Forecasts

• +/‐ 30 day search window
• 100 analog forecasts

http://www.esrl.noaa.gov/psd/forecasts/ reforecast/

Agricultural and Biological Engineering

Precipitation Forecast Skill

‐0.05 0.05 0.1 0.15 0.2 0.25 0.3 0.35 J F M A M J J A S O N D CRPSS

24‐hr 48‐hr 72‐hr 5‐day week‐1 week‐2

‐0.05 0.05 0.1 0.15 0.2 0.25 0.3 0.35 J F M A M J J A S O N D CRPSS

Day 1 Day 2 Day 3 Day 5 Day 7 0.2 0.4 0.6 0.8 1 1 2 3 4 5 6 Forecast Observation

F(x) x

[ ] dx

(x) F F(x) CRPS

2

• ∫

∞ ∞ −

− =

CRPSS: Continuous Ranked Probability Skill Score

Ref Forecast

CRPS CRPS 1 CRPSS − =

Agricultural and Biological Engineering

Monthly/Seasonal Time‐scale

• Forecast analogs using CFS retrospective

forecast archive http://cfs.ncep.noaa.gov/

– Week 2 – Monthly – Seasonal

• Climate‐based probability of exceedance

streamflow forecasts

Agricultural and Biological Engineering

Probability of Exceedance Streamflow Forecasts

Posterior probability of streamflow conditioned on predictor Streamflow Forecast Streamflow Exceedance Probability (%) Season Lagged Niñ03.4

Correlation of Streamflow w/ ENSO

Agricultural and Biological Engineering

Decision Support + Scenarios

• Inputs:

– Forecasted Demand – Forecasted Withdrawal

• Outputs:

– Optimized source‐water allocations based on preferences/constraints – End of year reservoir volume

Agricultural and Biological Engineering

Lessons Learned

• There is a learning curve associated with using

weather/climate datasets!!! (for hydrologists/engineers, at least…)

• Limited number of forecasted variables

archived in retrospective archives may limit usefulness

Agricultural and Biological Engineering

Relevance

• Tools/approaches that can easily be replicated

in other regions

– Analog forecasts – Exceedance streamflow forecasts