noaa’s snow climatology dataset & user perspectives
DESCRIPTION
NOAA’s Snow Climatology Dataset & User Perspectives. Richard R. Heim Jr. NOAA/NESDIS/National Climatic Data Center Asheville, North Carolina Timothy Kearns NOAA/National Weather Service WFO, Aberdeen, South Dakota - PowerPoint PPT PresentationTRANSCRIPT
National Climatic Data Center
NOAA’s Snow Climatology Dataset & User Perspectives
Richard R. Heim Jr.NOAA/NESDIS/National Climatic Data Center
Asheville, North Carolina
Timothy KearnsNOAA/National Weather Service WFO, Aberdeen, South Dakota
Snowfall Observations and Products in the 21st Century: Meeting the Needs of FEMA and the Climate Community
Estes Park, CO – 25-27 May 2011
National Climatic Data Center
Snow Climatology Background
Developed in late 1990s Snowfall & Snow Depth climatologies for COOP stations
Support NWS operations 1997 ESDIM grant
SnowClim generated
CD – flat ascii files
Support FEMA snow disaster declarations 2000 FEMA grant
online access – tabular format
Other applications media, monitoring, research
http://www.ncdc.noaa.gov/ussc/
National Climatic Data Center
Variables & Statistics Mean Quartiles* NY* Extremes* Prob* RP* NF*
Snowfall X X
Daily & multi-day SF X X X X X
Total SF (monthly, seasonal, annual)
X X X X X
Number of days with SF** X X X
Number of consecutive days with SF**
X X X X
First/last occurrence of SF** MD X X
Number of days with snow cover (SD)**
X X X
Daily SD X X X
Number of consecutive days with snow cover (SD)**
X X X X
Snow Cover (SD) X X
*Legend: Quartiles: first quartile, median (MD), third quartile Extremes: greatest, top/bottom 10 (for monthly SF), earliest/latest (for first/last occurrence of SF) Prob: probability of receiving daily SF in a month NF: number of years with SF or SD on that day meeting the SF or SD threshold; when used in conjunction with NY, daily probability or frequency of snow = NF / NY
NY: number of years with non-missing dataRP: return periods for 1-day, 2-day, 3-day, monthly SF
** meeting several specific thresholds
National Climatic Data Center
Variables & Statistics Mean Quartiles* NY* Extremes* Prob* RP* NF*
Snowfall X X
Daily & multi-day SF X X X X X
Total SF (monthly, seasonal, annual)
X X X X X
Number of days with SF** X X X
Number of consecutive days with SF**
X X X X
First/last occurrence of SF** MD X X
Number of days with snow cover (SD)**
X X X
Daily SD X X X
Number of consecutive days with snow cover (SD)**
X X X X
Snow Cover (SD) X X
*Legend: Quartiles: first quartile, median (MD), third quartile Extremes: greatest, top/bottom 10 (for monthly SF), earliest/latest (for first/last occurrence of SF) Prob: probability of receiving daily SF in a month NF: number of years with SF or SD on that day meeting the SF or SD threshold; when used in conjunction with NY, daily probability or frequency of snow = NF / NY
NY: number of years with non-missing dataRP: return periods for 1-day, 2-day, 3-day, monthly SF
** meeting several specific thresholds
FEMA Snow Disaster Declaration Statistics
National Climatic Data Center
Snow Climatology Processing
Stations analyzed All COOP stations analyzed – 24,040
TD-3200: Summary of the Day, 1948-present, operational QC TD-3206: Data Rescue, pre-1948, limited QC
A subset of stations made available Exclude stations will short records (< 15 years of non-missing data) Current and closed stations for historical climatologies 9099 stations for FEMA snow disaster declaration support Not all counties represented
The mix of data sources required stringent QC Accepted TD-3200/3206 QC Applied additional QC
National Climatic Data Center
Snow Climatology Processing Quality Control Applied
Accepted the ValHiDD (Reek et al., 1992) QC applied to TD-3200 data limits check, internal consistency checks, flatliner temperature check,
precipitation/snowfall/snow depth (PSFSD) relationship check, temperature spike check, multiple rule-group failures check, and failed fix check
In some PSFSD cases, ValHiDD could not identify which element should be corrected, so the values were flagged as suspect and not altered
Additional QC applied – data values changed if fail temporal checks (today’s SD compared to yesterday’s SD & SF) factor of 10 check for SF (if SF/P > 80, then SF=SF/10) hail check (nonzero SF set to zero if TMIN >= 40) nonzero SF set to missing if
SF > 0.4 but P = 0, or today’s P is missing
factor of 10 check for SD (SD divided by 10 or set to missing) zero SD set to missing if yesterday’s SD > 7 and today’s SF > 2 nonzero SD set to missing depending on SD, SF, &/or TMAX criteria SF & SD extremes check (based on state extremes from Ludlum [1982] & NCDC
data base extremes)
National Climatic Data Center
Snow Climatology Processing Quality Control Applied
Additional QC applied – data values not changed questionable SF values flagged but not changed if SF/P ratios were unusual questionable SD values flagged but not changed based on SD, SF, & TMAX criteria
QCI (Quality Control – Inventory – Metadata) Statistics Useful for selecting the best quality stations Data QC indicators – number of non-missing daily values read, flagged, corrected,
not corrected, failed QC (& percent of total for these indicators)
Data Inventory indicators – number of years in data base, complete months, number & percent daily values missing & processed, gap (break) info
Station Metadata indicators – number of location changes, ob time changes, observer changes (& these indicators scaled by number of years in metadata base)
National Climatic Data Center
Snow Climatology Processing Daily to Monthly Values – Tolerance for Missing Data
Total SF – zero tolerance for missing data means & other statistics computed from year-month sequential data year-month sequential monthly or seasonal or annual SF value not computed if
even one day was missing the six seasons (especially annual and August- July) have a greater chance of
experiencing missing data and, generally, will have fewer years with non-missing data when compared to the individual months
Median daily value for a month – zero tolerance Number of days with SF or SD – zero tolerance Consecutive days with SF or SD – zero tolerance Daily extreme, multiple-day extreme, date of occurrence
could tolerate up to 5 missing days in a month
Greatest 2-day & 3-day SF could tolerate up to 5 missing days in a month but if a 2-day or 3-day period had a missing day within that period, then that period
was omitted from the analysis
National Climatic Data Center
Snow Climatology Processing Daily to Monthly Values – The Effect of Missing Data
Impact: Very stringent QC Potentially “good” daily SF values may be
set to missing or changed Fewer monthly (& seasonal) total SF values available for the
climatologies
Therefore …
Snow Climatology monthly SF extremes (and other monthly statistics) may differ from monthly SF extremes from other sources
Snow Climatology daily SF extremes may differ from daily SF extremes from other sources
Question: When We Reprocess Later This Summer … Revise QC? – Revise criteria for computing monthly totals?
Breakout Session
NWS Perspective onNCDC Snow Climatology
- Surveyed 10 Snowy WFO’s- 8 WFO’s responded- 7 were not familiar with NCDC’s Snow Product- All 8 of the WFO’s found differences between
- WFO’s Database- NCDC’s Snow Climatology- ACIS Database
- WFO’s Reported Significant Differences with NCDC on Extreme Values
NWS Perspective onNCDC Snow Climatology
- Primary Reasons for Differences in Extreme Values NCDC versus NWS
- NCDC - Employs rigid Quality Control- Intolerant of Missing Data
- NWS- Limited Quality Control
NWS Perspective Who Uses Snow Data
• NWS’s Primary Snow Customers• Public• Press• Emergency Managers• City Planners (snow removal)• Farmers• Insurance Companies• Other Government Agencies
What Type of Data is Requested
• NWS’s Customers Requests
• Departure from Normal
• Extremes and where winter/month/day ranks
NWS Perspective onNCDC Snow Climatology
- Extreme Example of Differences Between NCDC Snow Climatology and NWS Local Database
Top Ten Snowiest YearsAugust - July
• Aberdeen’s Database
• 1 109.8 1937• 2 94.6 1897 • 3 79.3 2011 • 4 76.8 1994• 5 75.9 1997• 6 74.6 2001 • 7 74.5 1915 • 8 71.3 1907 • 9 68.0 1936• 10 67.5 1893 • POR 1893 - 2011
• NCDC’s Snow Climatology
• 1 52.8 1935• 2 39.2 1933• 3 29.5 1944• 4 27.5 1990• 5 24.7 1934• 6 24.4 1946• 7 23.4 1945• 8 19.9 1942• 9 19.8 1941• 10 19.1 1957• POR 1893 - 2006
NWS Perspective onNCDC Snow Climatology
• Discussion Points
• Where’s the Middle Ground?
• What do our Customers Want?
Breakout Session
National Climatic Data Center
Thank You!
Snow Climatology:http://www.ncdc.noaa.gov/ussc/
NCDC Climate Monitoring Branch Reports & Products:http://www.ncdc.noaa.gov/oa/climate/research/monitoring.html