GFE Diurnal Temperature Trend Smart Tool

1
GFE Diurnal Temperature TrendSmart Tool
Chris Sohl Norman OK
2
The Motivation

• Grids appear very accurate and detailed, but
• Important forecasts occasionally have large errors

10PM ASOS Temps SPS 82 OKC 79 HBR 78 GAG 77
3
The Motivation

• This error often results from cubic interpolation
  of only 2 T grids per 24 hours (i.e. MaxT and
  MinT)

4
The Motivation

• Linear interpolation has saw tooth appearance
• Cubic spline interpolation shifts the time of Max
  and Min temperatures
• Both methods too slow to warm temps in the day
  and cool in the evening

5
(No Transcript)
6
The Motivation

• Computed typical temperature changes at certain
  intervals of the day and adjusted those grids
  accordingly before interpolating
• Thus, why not develop an interpolation algorithm
  that performs this operation?

7
Methodology

• Used 8 years of 5 minute data for two Mesonet
  sites in Oklahoma
• Determined the average ratio (f) of the Min. to
  Max. temperature at each hour
• Diurnal temperature curves derived for each month
• Result DiurnalTfromMaxTMinT

8
Methodology

• J F M A M J J A S
  O N D GMT
• f 68, 81, 86, 89, 87, 89, 89, 86, 80,
  69, 62, 61, 0
• 53, 62, 65, 72, 72, 75, 75, 69, 57,
  51, 50, 50, 1
• 44, 50, 53, 54, 53, 55, 55, 50, 45,
  41, 42, 40, 2
• 37, 42, 44, 44, 42, 43, 42, 40, 37,
  35, 35, 34, 3
• 31, 36, 36, 36, 34, 35, 34, 33, 30,
  30, 30, 29, 4
• 26, 31, 30, 30, 27, 28, 28, 27, 24,
  25, 25, 25, 5
• 21, 25, 24, 25, 21, 22, 22, 22, 20,
  20, 20, 21, 6
• 17, 20, 18, 20, 16, 16, 17, 19, 16,
  17, 16, 18, 7
• 13, 16, 15, 15, 12, 11, 13, 14, 12,
  14, 13, 15, 8
• 10, 12, 11, 10, 7, 8, 8, 10, 9,
  11, 10, 11, 9
• 7, 9, 7, 6, 3, 3, 4, 6, 5,
  7, 7, 7, 10
• 5, 5, 3, 3, 0, 0, 1, 2, 2,
  4, 5, 5, 11
• 2, 2, 1, 0, 0, 1, 0, 0, 0,
  1, 1, 1, 12
• 0, 0, 0, 6, 12, 15, 13, 10, 3,
  0, 0, 0, 13
• 1, 2, 12, 24, 28, 31, 29, 28, 21,
  16, 7, 1, 14
• 15, 20, 30, 41, 44, 47, 46, 46, 41,
  37, 26, 19, 15
• 36, 41, 48, 56, 59, 60, 60, 62, 58,
  56, 47, 40, 16
• 56, 59, 63, 69, 71, 72, 74, 76, 72,
  71, 65, 60, 17

Python f-table
9
f-table
10
Features of DiurnalTfromMaxTMinT

• f-table used on Fcst MaxT and MinT to compute
  the hourly T that climatologically is the warmest
  or coldest for that time of year
• Diurnal curves factor in average wind, cloud
  cover and humidity

11
Tool Limitations

• Not for periods with a non-standard diurnal
  temperature trace (i.e. falling temps during the
  day or warming at night)
• Periods with non-standard trace can be deselected

12
Advantages

• An improved interpolation scheme over existing
  linear or cubic spline interpolation methods
• Tool based on meteorology (climatology) rather
  than a strictly mathematical derivation

13
Advantages

• Provides better first guess temperature trend
  which can be easily modified with the temporal
  tool
• Especially useful for days 3 through 7

14
Applying the Tool

• MaxT and MinT grids edited
• Smart Tool applied to those periods where a
  standard diurnal trace exists

15
(No Transcript)
16
Adjustments to diurnal curve easily made in the
temporal editor (EX faster than normal
cooling during evening hours on a calm,
clear night)
17
Examples of Non-standard diurnal trace
Inverted sharkfin shape of the climatological
trace
18
Max/Min points on curve in temporal editor are
not adjusted Can adjust temporal curve for
smaller edit area subset
19
Credits

• Cheryl Sharpe Concept Statistical
• Analysis of Climate
  Data
• Steve Nelson Smart Tool Coding
• Oklahoma Climatological Survey Provided
  climate data