ML17271A016

Meeting Slide for Tennessee Valley Authority (TVA) Hydrology Precipitation Model Topical Report Review Status

ML17271A016
Person / Time
Site:	Browns Ferry, Watts Bar, Sequoyah
Issue date:	09/28/2017
From:	Plant Licensing Branch II
To:
Hon A
References
Download: ML17271A016 (46)

Text

TVA Overall Basin Probable Maximum Precipitation (PMP) and Local Intense Precipitation Analysis Status of Staffs Review Topical Report TVA-NPG-AWA16 September 28, 2017

Status of Review TVA Overall Basin Probable Maximum Precipitation and Local Intense Precipitation Analysis (TVA-NPG-AWA16)

• TVA submitted a Topical Report (TR) in September 2016, which will be used for subsequent licensing actions for three operating TVA nuclear power plants.

• NRC staff requested information to facilitate review of the TVA TR.

- NRC staffs review of the TR is informed by previous site-specific probable maximum precipitation (ssPMP) reviews, including Flood Hazard Reevaluation Reports and Mitigating Strategies Assessments, observation and participation in other agency reviews, and extensive experience in meteorology.

• Audit meeting held in May 2017.

• Information Need responses received and reviewed.

• A number of key issue areas are resolved:

- Local and General storm transposition limits

- Dew point climatology smoothing

- Tropical PMP development (limited size of storm list)

- Storm selection screening

• Staff identified 3 Key Areas of Concern for concluding the review.

2

Presentation Overview

• Sensitivity Methodology

• Staffs Key Areas of Concern A. Dew Point

1. Storm Representative Dew Point Selection

2. Dew Point Climatology B. Tropical Storms

1. Tropical Storm Custom Transposition Limits C. Orographic Transposition Factor (OTF)

1. Use of OTF in TVA Zones 1-3

2. OTF Reduction for the Smethport, PA and Simpson, KY Storms

3. OTF Calculation Methodology

4. OTF Lower Limit 3

NRC Staff Review Evaluation and Key Areas of Concern 4

Sensitivity Analysis

• Staff performed confirmatory independent PMP calculations using input data from TVA.

• Staff conducted sensitivity analysis to determine the impact of staffs key concerns on PMP.

- Select basins used for sensitivity analysis are shown on subsequent slides.

5

Maps of TVA Basin Elevation and Zones Used in Topical Report 6

Maps of Select Basins Used in Sensitivity Analysis Norris: Cherokee:

Zones 2 & 3 Zones 3 & 4 Douglas: Fontana:

Zones 3 & 4 Zone 4 7

Maps of Select Basins Used in Sensitivity Analysis Hiwassee: Watts Bar:

Zone 4 Zones 2, 3, 4 Sequoyah: Browns Ferry:

Zones 2, 3, 4 Zones 1, 2, 3, 4 8

Key Area of Concern A - Dew Point Key Concern #A1 Storm Representative Dew Point Selection 9

Key Concern #A1: Storm Representative Dew Point Selection

• For several controlling storms, the licensees storm representative dew point timeframe (accounting for travel time) corresponds to a period after much of the storms rainfall had already occurred.

- Big Rapids, MI - Representative of the moisture feeding the last one-third (approximately) of the rainfall event.

- Warner Park, TN - Representative of the moisture conditions relevant to the very last portion of the rainfall event, after most of the rainfall has already occurred.

- Wellsville, NY - Representative of the moisture feeding the last one-third (approximately) of the rainfall event.

- Americus, GA - Not representative of the moisture feeding any of the rainfall event.

• Similar issue identified during previous NRC SSPMP reviews.

10

Key Concern #A1: Storm Representative Dew Point Selection Example for Warner Park, TN (controlling General storm)

Storm center location Distance: ~230 miles Avg. wind speed: 16mph Travel time: 22.5h Region where rainfall occurred Region where storm representative dew point is analyzed Source: TVA Topical Report 11

Key Concern #A1: Storm Representative Dew Point Selection

• Dew point (moisture) timeframe selected only contributes to last part of rainfall.

Example for Warner Park, TN (controlling General storm) 12

Key Concern #A1: Storm Representative Dew Point Selection

• Dew point (moisture) timeframe selected only contributes to last part of rainfall.

- Using an earlier timeframe, the storm representative dew point would be lower and result in a more accurate representation of the moisture source for the bulk of the rainfall.

Example for Warner Park, TN (controlling General storm) 13

Key Concern #A1: Storm Representative Dew Point Selection 14

Key Concern #A1: Storm Representative Dew Point Selection

• PMP sensitivity addressing Key Concern #A1:

- PMP is greater than or equal to licensees results when using more reasonable dew points for select controlling storms:

- Gridded General PMP increases by up to 6%.

- Gridded Tropical PMP increases by up to 10%.

15

Key Concern #A1: Storm Representative Dew Point Selection

% Difference in basin-averaged General PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 4% 4% 2% 1% 4% 3% 3% 2%

Sequoyah Watershed 4% 5% 2% 2% 4% 4% 3% 3%

Watts Bar Watershed 4% 5% 3% 2% 4% 4% 3% 3%

Norris Watershed 5% 5% 4% 2% 4% 4% 3% 3%

Cherokee Watershed 4% 5% 3% 2% 2% 2% 2% 2%

Douglas Watershed 5% 5% 4% 4% 4% 4% 4% 4%

Fontana Watershed 5% 5% 5% 5% 5% 5% 5% 5%

Hiwassee Watershed 5% 5% 5% 5% 5% 5% 5% 5%

% Difference in basin-averaged Tropical PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 0% 0% 0% 0% 1% 4% 5% 6%

Sequoyah Watershed 0% 0% 0% 0% 2% 5% 6% 6%

Watts Bar Watershed 0% 0% 0% 0% 2% 5% 5% 5%

Norris Watershed 0% 0% 0% 0% 1% 1% 1% 1%

Cherokee Watershed 0% 0% 0% 0% 4% 4% 5% 5%

Douglas Watershed 0% 0% 0% 0% 7% 8% 8% 8%

Fontana Watershed 5% 0% 0% 0% 5% 8% 8% 8%

Hiwassee Watershed 4% 0% 0% 0% 6% 8% 8% 8%

% Difference between staff and licensee PMP

-20% -10% 0 +10% +20%

16

Key Area of Concern A - Dew Point Key Concern #A2 Dew Point Climatology 17

Key Concern #A2: Dew Point Climatology

• The licensees dew point climatology appears to result in lower PMP values due to possible raw data and data processing issues.

• NRC staff compared the licensees annual maximum series (AMS) dew point data (based on NOAA TDL) with NRC-analyzed AMS dew point data (based on NOAA TD3505).

- NOAA TD3505 is subject to higher QA/QC than NOAA TDL.

• TDL is a near real-time and is not re-processed or revised.

• TD3505 is re-processed and revised.

- Staff observed several low or missing AMS values and the inclusion of duplicate years in the licensees AMS data.

- NRC AMS data generally contain higher dew points than licensee AMS data.

- The NRC-analyzed dew point climatology values are generally higher than the licensee-analyzed values.

- An Excel file was generated to highlight many significant differences in the AMS data.

• Similar issue identified during previous NRC SSPMP reviews.

18

Key Concern #A2: Dew Point Climatology

• AMS of dew point used to compute 100-y dew point climatology (example MEI station data below).

AWA Annual Maximum Dew Point YY MM DD HH Max Td ORNL Annual Maximum Dew Point YY MM DD HH Max Td ORNL-AWA Highlighted cells 1942 5 14 23 73.4 indicate large 1943 5 28 16 72.7 differences between 1944 5 29 23 71.3 NRC and TVA climatology.

AMS comparison of May climatology for Meridian, Mississippi (MEI) station Consistent bias seen 8005 1949 5 30 18 71 70.7 6 1949 5 30 20 70.7 0.0 16108 1950 5 3 9 70 70.2 6 1950 5 3 11 70.2 0.0 when comparing both 25520 1951 5 30 13 70 71.2 6 1951 5 30 15 71.2 0.0 AMS data and 34338 1952 5 31 23 -999 66 6 1952 5 17 17 71.5 5.5 resulting climatology.

43085 1953 5 31 10 69 69.8 4 1953 5 1 22 74.1 4.3 51858 1954 5 31 23 -999 71.5 6 1954 5 1 17 70.7 -0.8 60618 1955 5 31 23 -999 58 6 1955 5 23 15 74.1 16.1 69402 1956 5 31 23 -999 70.5 6 1956 5 31 23 70.2 -0.3 Higher Max Td 78152 86913 1957 1958 5

5 31 31 13 14 66 68 69.2 69.8 6

6 1957 1958 5

5 23 29 13 0

73.0 71.5 3.8 1.7 results in increased 95676 1959 5 31 17 72 73.8 6 1959 5 25 17 78.3 4.5 PMP.

19

Key Concern #A2: Dew Point Climatology Regional differences in NRC-TVA station-based dew point climatology (May; 6-, 12-, and 24-h) 6-h Dew Point Climatology Delta - May June 12-h Dew Point Climatology Delta - May June 24-h Dew Point Climatology Delta - May June

-5 F 0 +5 F -5 F 0 +5 F -5 F 0 +5 F Red: NRC station-based climatology higher Blue: TVA station-based climatology higher 20

Key Concern #A2: Dew Point Climatology

• PMP sensitivity addressing Key Concern #A2:

- Recalculated the dew point climatology for all storms using the TD3505 dataset (instead of TDL) and addressed data processing concerns.

- PMP is generally greater than licensees results:

- Gridded General PMP increases by up to 18%.

- Gridded Tropical PMP generally increases for longer duration and in the upper watershed.

21

Key Concern #A2: Dew Point Climatology

% Difference in basin-averaged General PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 12% 11% 14% 15% 13% 13% 12% 9%

Sequoyah Watershed 10% 10% 16% 16% 13% 12% 11% 10%

Watts Bar Watershed 8% 8% 14% 16% 12% 10% 10% 10%

Norris Watershed 2% 2% 3% 18% 9% 8% 7% 7%

Cherokee Watershed 4% 4% 7% 12% 8% 8% 7% 7%

Douglas Watershed 6% 6% 8% 11% 8% 8% 8% 8%

Fontana Watershed 6% 6% 6% 6% 6% 6% 6% 6%

Hiwassee Watershed 11% 11% 11% 11% 11% 11% 11% 11%

% Difference in site-specific Local PMP (1-mi2)

Site 1hr Browns Ferry Nuclear 6%

Sequoyah Nuclear -2%

Watts Bar Nuclear -1%

% Difference in basin-averaged Tropical PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 1% 1% 1% 1% 3% 4% 2% 2%

Sequoyah Watershed 1% 1% 1% 1% 5% 7% 5% 5%

Watts Bar Watershed 2% 2% 2% 2% 4% 5% 3% 3%

Norris Watershed 3% 3% 3% 3% 3% 3% 3% 3%

Cherokee Watershed 1% 1% 0% 0% 0% 0% 0% 0%

Douglas Watershed 3% 3% 2% 1% 8% 7% 8% 8%

Fontana Watershed 3% 1% -4% -5% 8% 10% 10% 9%

Hiwassee Watershed 4% 8% 0% 1% 20% 20% 21% 20%

% Difference between staff and licensee PMP 22

-20% -10% 0 +10% +20%

Summary for Key Area of Concern A - Dew Point

• For several controlling storms, the licensees storm representative dew point timeframe (accounting for travel time) corresponds to a period after much of the storms rainfall had already occurred.

• NRC staff compared the licensees annual maximum series (AMS) dew point data (based on NOAA TDL) with NRC-analyzed AMS dew point data (based on NOAA TD3505).

• The licensees dew point climatology appears to underestimate PMP values due to possible raw data and data processing issues.

23

Key Area of Concern B - Tropical Storms Key Concern #B1 Tropical Storm Custom Transposition Limits 24

Key Concern #B1: Tropical Storm Custom Transposition Limits

• The licensees use of Tropical Storm Remnant (TSR) L-Cv 0.24 contour to specify transposition limits does not have a physical basis.

- L-Cv is a statistical parameter defined by the linear moment of the annual maximum rainfall depths in the region.

- Staff is not aware of L-Cv contours being used in other PMP studies to specify transposition limits.

- This affects both Americus, GA and Larto Lake, LA storms.

25

Key Concern #B1: Tropical Storm Custom Transposition Limits Approximate 0.24 L-Cv contour 26 Source: MGS (2015) study for TVA

Key Concern #B1: Tropical Storm Custom Transposition Limits

• PMP sensitivity analysis addressing Key Concern #B1:

- Removing the TSR L-Cv 0.24 transposition limits for Americus, GA and Larto Lake, LA, results in significantly increased gridded Tropical PMP values in the upper TVA basin.

27

Key Concern #B1: Tropical Storm Custom Transposition Limits

% Difference in basin-averaged Tropical PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 0% 0% 0% 0% 0% 0% 3% 4%

Sequoyah Watershed 0% 0% 0% 0% 0% 0% 5% 7%

Watts Bar Watershed 0% 0% 0% 0% 0% 2% 7% 11%

Norris Watershed 0% 0% 0% 0% 14% 33% 35% 36%

Cherokee Watershed 0% 0% 0% 0% 6% 17% 18% 19%

Douglas Watershed 0% 0% 0% 0% 1% 5% 5% 5%

Fontana Watershed 2% 0% 0% 0% 0% 1% 1% 1%

Hiwassee Watershed 0% 0% 0% 0% 0% 0% 0% 0%

% Difference between staff and licensee PMP

-20% -10% 0 +10% +20%

28

Summary for Key Area of Concern B - Transposition Limits

• The use of the TSR L-Cv 0.24 contour is not well-justified and does not have a physical basis.

• The staff is unaware of this limit being used in previous ssPMP studies.

• Inclusion of additional storms within the region increases PMP for certain watersheds within the region.

Note: Moisture from two recent major tropical storm remnants (Harvey and Irma) have influenced precipitation in the TN Valley in 2017.

29

Key Area of Concern C - Orographic Transposition Factor Key Concern #C1 Use of OTF in TVA Zones 1-3 30

Key Concern #C1: Use of OTF in TVA Zones 1-3

• For the mountainous TVA Zone 4, there is significant OTF variation and the resulting pattern is consistent with expected precipitation patterns in complex terrain.

- The use of the OTF, with modification, in TVA Zone 4 appears adequate.

• Staff continues to have concerns with the use of OTF in Zones 1-3.

- For the relatively flat TVA Zones 1-3, OTF significantly reduces storm precipitation in some cases that do not have a clear, quantitative, physical basis and instead are related to variations in precipitation frequency contouring.

- Atlas 14 is a frequency-based statistical product developed over many years resulting in inconsistencies in statistical methodologies between NWS Atlas 14 geographical areas.

- This introduces non-physically related sources of uncertainty into the maps that are the basis of the OTF. This becomes most apparent in the application to non-orographic areas.

31

Key Concern #C1: Use of OTF in TVA Zones 1-3

• NRCs review demonstrates unexplained differences in OTF values in areas of similar topography.

Storm Rationale Observations Warner, OK The storm center location and TVA Zone 1 share similar orographic Avg. Zone 1 OTF: 0.80 (Example 1) characteristics (e.g., both locations exhibit minimal barriers, are of similar elevation, and are located at a similar distance-to-coast).

Fall River, KS The storm center location and TVA Zone 1 share similar orographic Avg. Zone 1 OTF: 0.75 (Example 2) characteristics (e.g., both locations exhibit minimal barriers, are of similar elevation, and are located at a similar distance-to-coast).

32

Key Concern #C1: Use of OTF in TVA Zones 1-3 OTF OTF varies more than

+/- 50% MTF, yet there is NOAA Atlas 14 1,000-year, 24-hr Precipitation Frequency greater uncertainty (Source: TVA Topical Report) in the OTF.

MTF Vol 2 +/- 10%

Vol 8 Different NOAA Atlas 14 Volumes. Volumes use Vol 9 different methods, resulting in different OTF values at boundaries.

Vol 11 Spatial anomaly resulting from using observations for statistical product.

33

Key Area of Concern C - Orographic Transposition Factor Key Concern #C2 OTF Reduction for the Smethport, PA and Simpson, KY Storms 34

Key Concern #C2: OTF Reduction for the Smethport, PA and Simpson, KY Storms

• The normalization method in the TR adjusted the Smethport and Simpson storms to reduce OTF values (and PMP) in a significant portion of the TVA Basin, including at TVA nuclear sites.

- The TR assumed that Smethport and Simpsons precipitation resulted from storms that were near maximum efficiency. The unadjusted transpositioned storms exceeded world record rainfall amounts in some locations.

- To address this issue, the TR normalized the OTF values throughout the entire basin for these storms based on the maximum gridded value.

35

Key Concern #C2: OTF Reduction for the Smethport, PA and Simpson, KY Storms Example OTF spatial distribution (Source: TVA Topical Report)

Location of highest OTF in TVA basin where OTF is artificially set = 1.00.

OTF for all other grids is reduced via normalization:

• Divide by 2.15 for Smethport

• Divide by 2.09 for Simpson

• No physical basis for applying these adjustments across the entire basin 36

Key Concern #C2: OTF Reduction for the Smethport, PA and Simpson, KY Storms

• PMP sensitivity addressing Key Concern #C2:

- Eliminated the use of OTF normalization for Smethport and Simpson.

- Watts Bar Nuclear: 61% increase in 1-h, 1-mi2 PMP.

- Sequoyah Nuclear: 61% increase in 1-h, 1-mi2 PMP.

- Browns Ferry Nuclear: no change (neither storm transpositioned to Browns Ferry).

37

Key Area of Concern C - Orographic Transposition Factor Key Concern #C3 OTF Calculation Methodology 38

Key Concern #C3: OTF Calculation Methodology

• TVA calculated OTF using a best-fit linear estimation based on the Atlas 14 10- to 1,000-year precipitation frequency values.

- A comparison of 100-year Atlas 14 values appears to provide a more stable approach than the best-fit linear estimation.

- Other PMP studies (such as the ongoing Colorado/New Mexico Regional Extreme Precipitation Study) have calculated the OTF using the 100-year precipitation frequency ratio rather than the linear regression approach.

• OTF sensitivity to duration has not been assessed.

- Use of a 6-h duration used for Local storms may produce different PMP depths than using a 1-h duration.

39

Key Concern #C3: OTF Calculation Methodology

• OTF values shown are non-normalized.

Using the 100-year ratio results in a final OTF spatial distribution that matches the underlying Atlas 14 PF spatial distribution. For example, the OTF rank and Atlas-14 rank (BFN>SQN>WBN) is consistent when using the 100-year ratio.

OTF linear best-fit method can be biased by intercept/fit.

OTF is sensitive to duration of the Atlas 14 precipitation depths. For this example, use of a 6-h duration for Local storms produces different PMP depths than using a 1-h duration.

40

Key Concern #C3: OTF Calculation Methodology

• PMP sensitivity addressing Key Concern #C3:

- Recalculated the OTF using the ratio between 100-year Atlas 14 precipitation frequency data rather than using linear extrapolation.

- PMP is generally greater than licensees results when calculating the OTF using the 100-year ratio.

- Gridded General PMP exhibits moderate spatial variation.

- Gridded Tropical PMP generally increases for longer duration.

41

Key Concern #C3: OTF Calculation Methodology

% Difference in basin-averaged General PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 1% 1% 1% 1% 2% 1% 1% -1%

Sequoyah Watershed -2% -2% -2% -2% 0% -1% -1% -2%

Watts Bar Watershed -1% -2% -2% -2% -1% -1% -1% -1%

Norris Watershed 1% -6% -5% -2% 2% 1% 1% 1%

Cherokee Watershed 7% 3% 0% 1% 1% 0% 0% 0%

Douglas Watershed 1% -1% -1% -2% -2% -2% -2% -2%

Fontana Watershed -4% -4% -4% -4% -4% -4% -4% -4%

Hiwassee Watershed -4% -4% -4% -4% -4% -4% -4% -4%

% Difference in site-specific Local PMP (1-mi2)

Site 1hr Browns Ferry Nuclear 6%

Sequoyah Nuclear 53% Note: no OTF normalization is used Watts Bar Nuclear 53% for this sensitivity analysis

% Difference in basin-averaged Tropical PMP Basin 6hr 12hr 18hr 24hr 48hr 72hr 96hr 120hr Browns Ferry Watershed 1% 1% 1% 1% 2% 8% 10% 11%

Sequoyah Watershed 2% 2% 2% 2% 3% 7% 8% 8%

Watts Bar Watershed 2% 2% 2% 2% 4% 7% 7% 7%

Norris Watershed -2% -2% -2% -2% -2% -2% -2% -2%

Cherokee Watershed 6% 6% 5% 6% 10% 10% 10% 10%

Douglas Watershed 3% 3% 3% 4% 12% 12% 12% 12%

Fontana Watershed 3% 3% 4% 5% 7% 7% 7% 7%

Hiwassee Watershed 2% 2% 3% 5% 8% 8% 8% 8%

% Difference between staff and licensee PMP 42

-20% -10% 0 +10% +20%

Key Area of Concern C - Orographic Transposition Factor Key Concern #C4 OTF Lower Limit 43

Key Concern #C4: OTF Lower Limit

• Application of a maximum OTF value of 1.5 without a corresponding minimum OTF value could lower PMP without an apparent physical basis.

• Since the OTF is a multiplicative adjustment factor, the use of a lower OTF limit should be compatible with the upper limit such that the lower limit is equal to the inverse of the upper limit (e.g., a 1.50 upper limit is compatible with a 0.67 lower limit).

- This ensures consistency when applying the OTF methodology between two locations, regardless of the source and target locations.

44

Summary for Key Area of Concern C - Orographic Transposition Factor

• Key Concern #C1: Use of OTF in TVA Zones 1-3

• Key Concern #C2: OTF Reduction for the Smethport, PA and Simpson, KY Storms

• Key Concern #C3: OTF Calculation Methodology

• Key Concern #C4: OTF Lower Limit 45

Summary and Path Forward

• Staff has conducted detailed technical reviews and sensitivity studies on key areas of TVAs Topical Report.

• A number of key issue areas have been resolved.

• Staff has three key remaining areas of concern:

- Dew Point

- Tropical Storms

- Orographic Transposition Factor (OTF)

• NRC staff will continue engaging with TVA to resolve the remaining concerns with the TR.

46