Mel Kunkel & Jen Pierce Boise State University Climatic Indices: Predictors of Idaho's Precipitation and Streamflow.

Mel Kunkel & Jen Pierce

Boise State University

Climatic Indices: Predictors of Idaho's Precipitation

and Streamflow

Outline

• Reconstructing Timing of Past Snowmelt

• Linking Climate Indices to the timing of snowmelt, snow accumulation and total precipitation

• The next steps…….

Reconstructing Timing of Past Snowmelt

• Purpose: Develop a repeatable technique for reconstructing past timing of snowmelt in Idaho to assess how past snowpacks have responded to oceanic and atmospheric influences --- this requires long term records.

Reconstructing Timing of Past Snowmelt

• SNOTEL/Snow Course:Daily SNOTEL and monthly snow course SWE data

is readily available from the NRCS online. NRCS snow course data (many from the 1930s) provides long-term monthly SWE readings while SNOTEL data provides detailed daily snow depth and SWE measurements, but in most cases these records only exist since the early 1980s.

There are over 350 SNOTEL and snow course (current and discontinued) sites located within the Idaho watershed.

Reconstructing Timing of Past Snowmelt

• SNOTEL/Snow Course:Recognizing that the NRCS emplaced SNOTEL

and snow course sites specifically to assist in forecasting the water supply of the western United States, we acknowledge that they may not be representative of the entire watershed.

Reconstructing Timing of Past Snowmelt

• Hydro-Climatic Data Network (HCDN): The U.S. Geological Survey (USGS) maintains the largest collection of stream gages in the United States; these gage records form the basis of the HCDN. Use of the HDCN for climate studies is well established (e.g. Dettinger and Cayan 1995; Cayan et al. 2001; Stewart et al. 2005; Jefferson et al. 2006; Barnett et al. 2008; Levi 2008).

• Since HCDN streamflow records are relatively free of anthropogenic influences (Slack and Landwehr 1992), variations in streamflow in the HCDN basins are likely caused by changes in climate or basin characteristics rather than dam-related flow regulation or other anthropogenically-induced change

Reconstructing Timing of Past Snowmelt

• Selecting suitable paired stream gage and SNOTEL site was critical to accurately reconstructing the timing of past snowmelt.

• We selected USGS gage stations and SNOTEL sites by spatially overlaying gage positions, SNOTEL positions, and stream/river information. We then identified gage sites and SNOTEL sites that were within the same drainage basin, and relatively close together.

Reconstructing Timing of Past Snowmelt

• Through this selection process, we identified 28 unregulated streams as candidates for use in extending the SNOTEL data sets.

• Other unregulated streams with gage sites lacked SNOTEL sites within their immediate drainages and were not used for this study.

Reconstructing Timing of Past Snowmelt

• Conducted a Short Term Fourier Transform on selected site discharge.

• STFT transforms the annual discharge data into a spectral signal that varies in shape and magnitude over time, reflecting changes in discharge.

• We developed procedures to use the results of the STFT to reconstruct the timing of snowmelt based on the temporal distance of final snowmelt from existing SNOTEL data from the spectral peak (peak spring discharge).

An example of a single spectral peak where the final snowmelt date always appears two timing points right of the spectral peak. (Trinity Mountain SNOTEL and Boise River)

An example of twin spectral. The timing point for this site always reads three timing points to the left of the left hand center of timing. (Meadow Lake SNOTEL and Lemhi River)

Reconstructing Timing of Past Snowmelt

Reconstructing Timing of Past Snowmelt

Reconstructing Timing of Past Snowmelt

• Results:• Reconstructions of snowmelt dates from these 19

sites extend records of the estimated timing of snowmelt back 42–97 years, >1100 reconstructions.

• None of the 419 compared reconstructed final snowmelt dates deviated from the actual snowmelt date by more 4 days ∼95% of the time and ±7 days 100%

• Assuming that the relationship between discharge and timing of snowmelt established using known data holds true for the earlier part of the record, we therefore assume reconstructions of the timing of snowmelt using this method are also within ±7 days.

Linking Climate Indices to the timing of snowmelt, snow accumulation and

total precipitation

Linking Climate Indices

• Use past records of the timing of snowmelt, max snow accumulation, dates of max accumulation and total precipitation

• Indices of interest– ENSO (TNI/MEI)– Arctic Oscillation (AO)– Pacific Decadal Oscillation (PDO)– And others

Linking Climate Indices

• 1999 – La Nina and Positive AO

Linking Climate IndicesMax SWE

Total Precip

Max SWE date

FinalSnowmelt

date

Date more than1.5 weeks later than normal

Date more than 1.5 weeks earlierthan normal

1999

10% or greaterabove normal

10% or greaterbelow normal

Linking Climate Indices

• 1999 – Greater than normal precipitation Later than normal final

snowmelt La Nina and Positive AO

• 1987 – El Nino and Negative AO

Linking Climate IndicesMax SWE

Total Precip

Max SWE date

FinalSnowmelt

date

Date more than1.5 weeks later than normal

Date more than 1.5 weeks earlierthan normal

1987

10% or greaterabove normal

10% or greaterbelow normal

Linking Climate Indices

• 1999 – Greater than normal precipitation Later than normal final

snowmelt La Nina and Positive AO

• 1987 – Lower than normal precipitation Earlier than normal final

snowmelt El Nino and Negative AO

• 1991 – El Nino and Positive AO

Linking Climate IndicesMax SWE

Total Precip

Max SWE date

FinalSnowmelt

date

Date more than1.5 weeks later than normal

Date more than 1.5 weeks earlierthan normal

1991

10% or greaterabove normal

10% or greaterbelow normal

Linking Climate Indices

• 1999 – Greater than normal precipitation Later than normal final snowmelt La Nina and Positive AO

• 1987 – Lower than normal precipitation Earlier than normal final snowmelt El Nino and Negative AO

• 1991 – Regional variations in precipitation Regional variation in final

snowmeltEl Nino and Positive AO

• 1998 – La Nina and Negative AO

Linking Climate IndicesMax SWE

Total Precip

Max SWE date

FinalSnowmelt

date

Date more than1.5 weeks later than normal

Date more than 1.5 weeks earlierthan normal

1998

10% or greaterabove normal

10% or greaterbelow normal

Linking Climate Indices• 1999 – Greater than normal precipitation

Later than normal final snowmelt La Nina and Positive AO

• 1987 – Lower than normal precipitation Earlier than normal final snowmelt El Nino and Negative AO

• 1991 – Regional variations in precipitation Regional variation in final snowmelt

El Nino and Positive AO• 1998 – Regional variations in precipitation

Regional variation in final snowmelt La Nina and Negative AO

The Next Step……

• Continue to use developed Empirical Rules of Thumb on how regional snow and precipitation responds under similar combined conditions to enhance understanding of physical processes driving Idaho weather

• Continued development of mixed statistical and physically based models for Idaho’s regional and watershed level snow, precipitation and streamflow

Thanks!!