Derived Values
1. General
a) Analysis techniques will be archived at each change so that
the analyses completed with a particular technique can be reconstructed
at any point.
b) Field versions of software will be maintained at a central
location for testing or trouble shooting.
c) All analysis to arrive at derived values will use standard
statistical techniques reviewed by the Design Team.
2. Determination of Central Tendency and Totals.
a) Central Tendencies will include means, medians, modes, and
the special case of means known as NORMALS. Normals should correspond
to those produced by NCDC for the CLIM 81 publication or an estimate
of NORMAL that is a calculation of a standard 30 year period average.
b) Record lengths for NORMALs estimation will be:
1) Minimum 25 years in length with no adjustment.
2) Minimum 20 years in length with no adjustment based on local
review.
3) Minimum 10 years of record with weighting based on statistical
comparison of short term record to a nearby station that meets
the minimum record length requirement. The short term record of
the site will then be adjusted to the long term record based on
the weighting factor.
c) NORMALS, NORMAL estimates, medians, and modes for the standard
30 year period or an estimated value for the 30 year period are
necessary.
d) Calculation of totals, i.e. total monthly precipitation, will
follow the same rules for length of record.
e) Missing data tolerances for calculations are established for
precipitation and temperature at this time.
1) Temperature values used in calculations require a minimum of
25 values. With local analysis determining whether 20 values is
sufficient.
2) Precipitation values used in calculations require a minimum
of 21 values. With local analysis determining whether 20 values
is sufficient.
3. Frequency Distribution and Probability
a) Frequency analysis will use the appropriate distribution for
the element and site location.
b) Where appropriate technique for distribution evaluation has
not been determined and/or the appropriate distribution has not
been determined the Log Pearson Type III analysis will be used.
c) Values are required for the following minimum probability of
exceedance:
0.001 ( 0.1%)
0.002 ( 0.2%)
0.005 ( 0.5%)
0.01 ( 1%)
0.02 ( 2%)
0.04 ( 4%)
0.05 ( 5%)
0.10 (10%)
0.20 (20%)
0.30 (30%)
0.50 (50%)
0.70 (70%)
0.80 (80%)
0.90 (90%)
0.95 (95%)
0.99 (99%)
1.00 (100%)
d) Frequency analysis is required for TD3200 stations, TD3220,
TD9641, and those stations where NORMALs have been determined.
4. Extremes and Updates
a) Extremes for TD3220 Monthly Values including dates
b) Extremes for TD3200 elements (Summary of the Day)
c) Provide 'near real-time' upgrades to P.O.R. Data
1) Delivery to Portland or Field Office
2) Value-added Products - Climatic Analysis