PDF Print Pages to print Enter page numbers and/or page ranges separated by commas. For example, 1,3,5-12.After downloading, print the document using a PDF reader (e.g. Adobe Reader). Download electronic document Show annotations interval). The following conversions are currently supported, with a description of the conversion process. Refer to the command parameter reference for an explanation of parameters.
The conversion from daily and monthly interval to yearly interval (for ACCM and MEAN) utilizes a simpler algorithm. Irregular Time Series to Regular Time Series An irregular time series
can be converted to a regular time series. The ability to change from an irregular or regular time series to an irregular time series currently is not implemented. Missing data is handled
in different ways depending on the old and new time scales. Each of the follow examples demonstrates how missing data is interpreted. The following conversion combinations are allowed.
Small Interval ACCM to Large Interval ACCM When converting from small interval accumulated data to large interval accumulated data, values from the old time series are summed for the
new interval-ending date/time from the values in the old intervals prior to this date/time. The following illustrates the conversion from NHour to NHour (1Hour to 3Hour example): Day
1, Hour 0 (A) Day 1, Hour 1 (B) Day 1, Hour 2 (Missing) Day 1, Hour 3 (C) Day 1, Hour 4 (Missing) Day 1, Hour 5 (Missing) Day 1, Hour 6 (Missing) Day 1, Hour 0 =A Day 1, Hour 3 =B+C
Day 1, Hour 6 = Missing 81 Command Reference – ChangeInterval() -1
ChangeInterval() Command TSTool Documentation Large Interval ACCM to Small Interval ACCM When converting from large interval accumulated data to small interval accumulated data, values
from the old time series are equally divided by the number of intervals prior to this date/time in the new time series since the previous non-missing data. The following illustrates
the conversion from NHour to NHour (3Hour to 1Hour example): Day 1, Hour 0 (A) Day 1, Hour 3 (B) Day 1, Hour 6 (Missing) Day 1, Hour 9 (C) Day 1, Hour 0 =A Day 1, Hour 1 =B/3 Day 1,
Hour 2 =B/3 Day 1, Hour 3 =B/3 Day 1, Hour 4 =C/6 Day 1, Hour 5 =C/6 Day 1, Hour 6 =C/6 Day 1, Hour 7 =C/6 Day 1, Hour 8 =C/6 Day 1, Hour 9 =C/6 Small Interval MEAN or INST to Large
Interval MEAN When converting from instantaneous or mean data to mean data, mean values are calculated for the new interval-ending date/time from the values in the old intervals prior
to this date/time. The following illustrates the conversion from NHour to NHour (1Hour to 3Hour example): Day 1, Hour 0 (A) Day 1, Hour 1 (B) Day 1, Hour 2 (Missing) Day 1, Hour 3 (C)
Day 1, Hour 4 (Missing) Day 1, Hour 5 (Missing) Day 1, Hour 6 (Missing) Day 1, Hour 0 =A Day 1, Hour 3 =(B+C)/2 Day 2, Hour 6 = Missing Large Interval MEAN or INST to Small Interval
MEAN When converting from large interval mean or instantaneous data to small interval mean data, values from the old time series are copied to the new interval-ending date/time time
series. The following illustrates the conversion from NHour to NHour (3Hour to 1Hour example): Day 1, Hour 0 (A) Day 1, Hour 3 (B) Day 1, Hour 6 (Missing) Day 1, Hour 9 (C) Day 1, Hour
0 =A Day 1, Hour 1 =B Day 1, Hour 2 =B Day 1, Hour 3 =B Day 1, Hour 4 =C Day 1, Hour 5 =C Day 1, Hour 6 =C Day 1, Hour 7 =C Day 1, Hour 8 =C Day 1, Hour 9 =C Small Interval INST to Large
Interval INST When converting from small interval instantaneous data to large interval instantaneous data, the data is copied directly from the old time series when available. If the
data is missing, the most recent previous valid data is used. The following illustrates the conversion from NHour to NHour (1Hour to 3Hour example): Command Reference – ChangeInterval()
-2 82
TSTool Documentation ChangeInterval() Command Day 1, Hour 0 (A) Day 1, Hour 1 (B) Day 1, Hour 2 (Missing) Day 1, Hour 3 (C) Day 1, Hour 4 (Missing) Day 1, Hour 5 (D) Day 1, Hour 6 (Missing)
Day 1, Hour 7 (E) Day 1, Hour 8 (F) Day 1, Hour 0 =A Day 1, Hour 3 =C Day 1, Hour 6 =D Large Interval INST to Small Interval INST When converting from large interval instantaneous data
to small interval instantaneous data, values from the old time series are linearly interpolated to calculate values for the new time series. The following illustrates the conversion
from NHour to NHour (3Hour to 1Hour example): Day 1, Hour 0 (A) Day 1, Hour 3 (B) Day 1, Hour 6 (Missing) Day 1, Hour 9 (C) Day 1, Hour 0 =A Day 1, Hour 1 =A+ (B-A)* (1/3) Day 1, Hour
2 =A+ (B-A)* (2/3) Day 1, Hour 3 =B Day 1, Hour 4 =B+ (C-B)* (1/6) Day 1, Hour 5 = B+ (C-B)* (2/6) Day 1, Hour 6 = B+ (C-B)* (3/6) Day 1, Hour 7 = B+ (C-B)* (4/6) Day 1, Hour 8 = B+
(C-B)* (5/6) Day 1, Hour 9 =C Regular Time Series to Regular Time Series ACCM (Accumulation) to ACCM (Accumulation) Small Interval ACCM (Accumulation) to Large Interval ACCM (Accumulation)
Changing the interval for small interval accumulated data to large interval accumulated data involves summing the small interval data values for the period that overlaps the large interval.
Accumulated data have a timestamp corresponding to the interval-end for the accumulation. Conversions involving time intervals that have zero values (e.g., Hour 0, Minute 0) result in
a perceived shift in time because the zero occurs on the boundary between larger intervals. The following examples illustrate the accumulation for common cases. In cases where an accumulation
jumps over two or more interval categories (e.g., minute to day), the accumulation occurs as if the two intermediate accumulations occurred in succession. In the following examples,
the general representation is shown first, followed by an example where appropriate. The following illustrates the conversion from NHour to Day (6Hour to Day example, i equals the hour
multiplier): Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 2, Hour 0 Day 1, Hour 0 Day 1, Hour 6 (A) Day 1, Hour 12 (B) Day 1, Hour 18 (C) Day 2, Hour 0 (D) Day 1 accumulation
(A+B+C+D) 83 Command Reference – ChangeInterval() -3
ChangeInterval() Command TSTool Documentation The following illustrates the conversion from NDay to Month (example for a month with 30 days): Month 1, Day 1 (A1) … … Month 1, Day 30
(A30) Month 1 accumulation (A1 + … + A30) Large Interval ACCM (Accumulation) to Small Interval ACCM (Accumulation) Changing from large interval accumulation data to small interval mean
data involves dividing each accumulated value by the number of new values for that same period of record. The following illustrates the conversion from Day to 6Hour (Day to 6Hour example,
i equals the hour multiplier): Day 1 accumulate (A) Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 1, Hour 0 = A/4 Day 1, Hour 6 = A/4 Day 1, Hour 12 = A/4 Day 1, Hour
18 = A/4 ACCM (Accumulation) to INST (Instantaneous) Accumulated to instantaneous is not currently supported. ACCM (Accumulation) to MEAN Small Interval ACCM to Large Interval MEAN See
Small Interval INST (Instantaneous) to Large Interval MEAN. Interval ACCM to Same Interval MEAN Changing the interval from accumulation data to the same interval mean data involves copying
the data from the old time series to the new time series (no changes to date values occur). The following illustrates the conversion from 6Hour to 6Hour (6Hour to 6Hour example, i equals
the hour multiplier): Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 1, Hour 0 (A) Day 1, Hour 6 (B) Day 1, Hour 12 (C) Day 1, Hour 18 (D) Day 1, Hour 0 =A Day 1, Hour
6 =B Day 1, Hour 12 =C Day 1, Hour 18 =D Command Reference – ChangeInterval() -4 84
TSTool Documentation ChangeInterval() Command Large Interval ACCM to Small Interval MEAN See Large Interval ACCM to Small Interval ACCM. INST (Instantaneous) to INST (Instantaneous)
Small Interval INST (Instantaneous) to Large Interval INST (Instantaneous) Changing the interval for small interval instantaneous data to large interval instantaneous data involves assigning
each date in the new time series a value from the corresponding date in the old time series. The HandleMissingInputHow parameter indicates how to interpret a missing value in the old
time series. HandleMissingInputHow=KeepMissing will simply assign a missing value for that date/time. HandleMissingInputHow=SetToZero will set the value to 0. Repeat fills the date with
data from the last non-missing value. Interpolation and using a non-missing future value may be added in the future. A special case is the ability to compute a statistic from the sample
of values from the input time series, using the Statistic parameter. For example, instantaneous 5 minute temperature data can be converted to 1 day maximum values. In this case, each
1 day sample of values from the input time series is used to compute the statistic. The initial handling of missing data described above is supported and additionally the AllowMissingCount
parameter is recognized to control computation of the statistic. The following illustrates the conversion from NHour to Day (6Hour to Day example where HandleMissingInputHow = Repeat,
i equals the hour multiplier): Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 1, Hour 0 (A) Day 1, Hour 6 (B)
Day 1, Hour 12 (C) Day 1, Hour 18 (D) Missing data Day 1, Hour 6 (E) Day 1, Hour 12 (F) Day 1, Hour 18 (G) Day 1 instantaneous = A Day 2 instantaneous = D Large Interval INST (Instantaneous)
to Small Interval INST (Instantaneous) Small interval instantaneous data is created from larger interval instantaneous data by linearly interpolating between the previous and current
large interval data to fill each value in the new time series during that same period of time. If the value in the old time series is missing, the method specified by the user in the
HandleMissingInputHow parameter is used. 85 Command Reference – ChangeInterval() -5
ChangeInterval() Command TSTool Documentation The following illustrates the conversion from Day to NHour (Day to 6Hour example, i equals the hour multiplier): Day 1 instantaneous (A)
Day 2 instantaneous (B) Day 3 … Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 2, Hour 0 Day 2, Hour i Day 2, Hour 2i Day 2, Hour 3i Day 1, Hour 0 =A Day 1, Hour 6 =A+
(B-A)* (6/24) Day 1, Hour 12 =A+ (B-A)* (12/24) Day 1, Hour 18 =A+ (B-A)* (18/24) Day 2, Hour 0 =B Day 2, Hour 6 … Day 2, Hour 12 … Day 2, Hour 18 … These values are an interpolated
value between the Day 1 instantaneous value and the Day 2 instantaneous value using a time of 24 hours. These values are an interpolated value between the Day 2 instantaneous value and
the Day 3 instantaneous value using a time of 24 hours. In the future, the ability to repeat input values may be added. INST (Instantaneous) to ACCM (Accumulation) Instantaneous to accumulated
is not currently supported. INST (Instantaneous) to MEAN Small Interval INST (Instantaneous) to Large Interval MEAN Changing from small interval instantaneous data to large interval
mean data involves adding together all the values from the small interval time series over the larger interval for the corresponding time period and then dividing by the number of data
values used within this calculation. As in other conversions, HandleMissingInputHow is first used to interpret missing data. If HandleEndpointHow = AverageEndpoints, the values at each
end of the interval are averaged for minute and hour inputs (the parameter does not apply to day, month or year input). The following illustrates the conversion from NHour to Day (6Hour
to Day example with HandleEndpointHow = IncludeFirstOnly, i equals the hour multiplier): Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 2, Hour 0 Day 2, Hour i Day 2,
Hour 2i Day 2, Hour 3i Day 1, Hour 0 Day 1, Hour 6 Day 1, Hour 12 Day 1, Hour 18 Day 2, Hour 0 Day 2, Hour 6 Day 2, Hour 12 Day 2, Hour 18 Value A B C D E F G H Day 1 mean= (A+B+C+D)/4
Day 2 mean=(E+F+G+H)/4 Command Reference – ChangeInterval() -6 86
TSTool Documentation ChangeInterval() Command The following illustrates the conversion from NHour to Day (6Hour to Day example with HandleEndpointHow = AverageEndpoints, i equals the
hour multiplier): Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 2, Hour 0 Day 2, Hour i Day 2, Hour 2i Day 2, Hour 3i Day 1, Hour 0 Day 1, Hour 6 Day 1, Hour 12 Day 1,
Hour 18 Day 2, Hour 0 Day 2, Hour 6 Day 2, Hour 12 Day 2, Hour 18 Value A B C D E F G H I Day 1 mean= ((A+E)/2 +B+C+D) /4 Day 2 mean=((E+I)/2+F+G+H) /4 Interval INST (Instantaneous)
to Same Interval MEAN If OutputFillMethod = Interpolate, see Large Interval INST (Instantaneous) to Small Interval INST (Instantaneous). Otherwise, the values are duplicated from the
old time series directly to the new time series. The following illustrates the conversion from 6Hour to 6Hour (6Hour to 6Hour example with OutputFillMethod = Repeat, i equals the hour
multiplier): Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 2, Hour 0 Day 1, Hour 0 (A) Day 1, Hour 6 (B) Day 1, Hour 12 (Missing) Day 1, Hour 18 (D) Day 2, Hour 0 (E)
Day 1, Hour 0 =A Day 1, Hour 6 =B Day 1, Hour 12 =B Day 1 Hour 18 =D Day 2, Hour 0 =E Large Interval INST (Instantaneous) to Small Interval MEAN If the OutputFillMethod = Interpolate,
see Large Interval INST (Instantaneous) to Small Interval INST (Instantaneous). The time series are handled in the same way. Otherwise, the values are duplicated from the old time series
directly to the new time series. The following illustrates the conversion from Day to 6Hour (Day to 6Hour example with OutputFillMethod = Repeat, i equals the hour multiplier): Day 1
instantaneous = A Day 1, Hour 0 Day 1, Hour i Day 1, Hour 2i Day 1, Hour 3i Day 1, Hour 0 =A Day 1, Hour 6 =A Day 1, Hour 12 =A Day 1, Hour 18 =A Each of these values is equal to the
instantaneous value for that day. 87 Command Reference – ChangeInterval() -7
ChangeInterval() Command TSTool Documentation MEAN to MEAN Small Interval MEAN to Large Interval MEAN See Small Interval INST (Instantaneous) to Large Interval MEAN. Large Interval MEAN
to Small Interval MEAN Changing from large interval mean data to small interval mean data involves copying values from the old time series into the new time series for that same period
of record. The following illustrates the conversion from Month to Day (Example for a month with 30): Month Mean (A) Day 1 =A Day 2 =A … Day 30 =A MEAN to ACCM (Accumulation) Small Interval
MEAN to Large Interval ACCM (Accumulation) See Small Interval INST (Instantaneous) to Large Interval MEAN. Interval MEAN to Same Interval ACCM (Accumulation) See Interval ACCM to Same
Interval MEAN. Large Interval MEAN to Small Interval ACCM (Accumulation) See Large Interval ACCM to Small Interval ACCM. MEAN to INST (Instantaneous) Small Interval MEAN to Large Interval
INST (Instantaneous) Not currently supported. Interval MEAN to Same Interval INST (Instantaneous) Not currently supported. The data can be treated equivalently by most commands. Large
Interval MEAN to Small Interval INST (Instantaneous) Changing the interval for large interval mean to small interval instantaneous data involves calculating a value for each new interval
based on trends found in the mean data. This approach has been adapted from the NWSRFS CHANGE-T operation (see http://www.nws.noaa.gov/oh/hrl/nwsrfs/users_manual/part5/_pdf/533changet.pdf).
The following example demonstrates how the data is converted from the old interval to the new interval. A general representation is shown first followed by an example. Command Reference
– ChangeInterval() -8 88
TSTool Documentation ChangeInterval() Command The following illustrates the conversion from Day to NHour (Day to 6Hour example, i equals the hour multiplier): Day 1 mean Day 2 mean Day
1 Day 1 Day 1 Day 1 Day 1 Ending endpoint Starting Int i Int 2i Int 3i Day 2 Starting endpoint endpoint Day 1 Day 1 Day 1 Day 1 Day 2 Hour 0 Hour 6 Hour 12 Hour 18 Hour 0 A B C D E In
computing instantaneous values, the volume of the original mean time series needs to be maintained. However, the value of the instantaneous endpoints affects the calculated mean value
for the previous and subsequent long intervals, since the mean over a longer interval uses both endpoints in its calculation. In the above example, Averagenew Day 1 = ((A+E)/2 +B+C+D)
/4 As a result, an iterative technique is required to adjust the initially computed instantaneous values to produce a time series with a volume that is within a specified tolerance of
the input mean volume for each time step. The following paragraphs describe how the initial instantaneous time series values are computed, followed by a description of the volume adjustment.
Initial Instantaneous Time Series Calculations Prior to converting from a large interval mean to small interval instantaneous, special cases are handled associated with missing data:
• Missing data is initially converted using the method specified by the user in the HandleMissingInputHow parameter. • If the current input value is still missing, the instantaneous
time series is also filled with missing data for each interval that falls in the larger interval. • If the previous or next mean values are missing, the current mean value for that interval
is copied directly to the instantaneous time series. The output instantaneous values for each input interval are computed using the current, next, and previous mean values. All three
values are useful because together they indicate whether the current value is part of a continuous rise or fall, a peak or trough or simply a continuation of a steady value. These conditions
are illustrated in the following figure. 89 Command Reference – ChangeInterval() -9
ChangeInterval() Command TSTool Documentation ChangeInterval_SQMEPic Mean data illustration • The first condition that may exist is a peak or trough. A peak exists when the current value
is greater than the previous and next values. A trough is when the current value is less than the next and previous values. 1. In this case, an instantaneous peak (or trough) is calculated.
Referring to the above illustration, the magnitude of the peak is calculated by adding (or subtracting for a trough) ¼ (a+b)/2 to the current mean. 2. The time of the instantaneous peak
is initially set to the start date/time of the current interval then shifted forward in time using the following calculation. The number of instantaneous intervals per larger interval
is multiplied by b/(a+b). That result is added to the start date/time. The time of the instantaneous peak will not necessarily correspond to the output interval. 3. The value for the
starting endpoint of the interval is set to the current value minus ¼ a. 4. The value for the ending endpoint of the interval is set to the current value minus ¼ b. 5. The remaining
intermediate instantaneous values for the interval are linearly interpolated between the peak (or trough) and both endpoints. • The second condition that may exist is a continuous rise
or fall. A continuous rise or fall exists when the current value is between the previous and next values. 1. In the case of a continuous rise, the starting endpoint of the interval is
set to the current value minus ¼ c (again using the above illustration). In the case of a continuous fall, ¼c is added to the current value. Command Reference – ChangeInterval() -10
90
TSTool Documentation ChangeInterval() Command 2. The ending endpoint of the interval of a continuous rise is set to the current value plus¼ c (minus ¼ c for a continuous fall) 3. The
remaining intermediate instantaneous values are calculated based on the following. a. The difference between the starting and ending endpoints is computed. b. The values c and a are
calculated. The ratio of mean differences is computed: If c > a, the mean ratio = c /a. If a > c, the mean ratio = a /c. c. If c is less than a, then the intermediate instantaneous values
are computed by adding small but increasing increments to the starting endpoint until the last point of the interval is reached. If a is less than c then the output values are computed
by subtracting small but increasing increments to the last endpoint until the first point of the interval is reached. For intermediate interval n from the appropriate endpoint: Incrementn
= (1/([number intervals] – n) * [endpoint difference] * [mean ratio]) Instantaneous Time Series Volume Adjustment After the instantaneous values are estimated using the above set of
rules, they are adjusted so that the volume over each interval is within a specified tolerance of the input mean volume for each time step. This tolerance is specified with the Tolerance
parameter. The volume calculation for each interval uses the average of the first and last endpoint. For each time step of the original mean time series, the error of each original interval
is computed as: ([mean of current instantaneous values] – [original mean]) /[original mean] If this volume error is within the tolerance, no adjustment is made for that time interval.
If the error is larger than the tolerance, the ratio: [original mean] /[mean of current instantaneous values] is computed. For the intermediate instantaneous values, the instantaneous
values are adjusted by multiplying each value by this ratio. For the instantaneous endpoint values, since these values affect the mean value of the current and the previous or following
time intervals, the endpoint values are not adjusted to the above calculated ratio. Instead, the new endpoint value is computed as the average of the original endpoint value and the
original endpoint value multiplied by the above ratio. The volume error is checked for each original time step. If any adjustments were made, the process is repeated, up to 15 iterations.
If the adjustment technique cannot adjust the instantaneous time series such that the corresponding mean volume is within the specified tolerance of the input mean volume within 15 iterations,
a warning will be written to the log file. The following dialog is used to edit the command and illustrates the syntax for the command. This example is converting a monthly volume time
series to annual water year (October to September) volumes. 91 Command Reference – ChangeInterval() -11
ChangeInterval() Command TSTool Documentation ChangeInterval ChangeInterval() Command Editor The command syntax is as follows: ChangeInterval(Parameter=Value,…) The following older command
syntax is updated to the above syntax when a command file is read: TS Alias = ChangeInterval(Parameter=Value,…) Command Parameters Parameter Description Default TSList Indicates the
list of time series to be processed, one of: • AllMatchingTSID – all time series that match the TSID (single TSID or TSID with wildcards) • AllTS – all time series generated before the
command AllTS Command Reference – ChangeInterval() -12 92
TSTool Documentation ChangeInterval() Command Parameter Description Default • EnsembleID – all time series in the ensemble • FirstMatchingTSID – the first time series that matches the
TSID (single TSID or TSID with wildcards) • LastMatchingTSID – the last time series that matches the TSID (single TSID or TSID with wildcards) • SelectedTS – the time series selected
with the SelectTimeSeries() command TSID The time series identifier or alias for the time series to be processed, using the * wildcard character to match multiple time series. Required
if TSList=*TSID. EnsembleID The ensemble to be processed, if processing an ensemble. Required if TSList= EnsembleID. Alias The alias to assign to the time series, as a literal string
or using the special formatting characters listed by the command editor. The alias is a short identifier used by other commands to locate time series for processing, as an alternative
to the time series identifier (TSID). None – must be specified. NewInterval The data interval for the new time series, from the provided choices. For example: 6Hour, Day, Month, Year.
None – must be specified. OldTimeScale The time scale for the original time series, one of: ACCM – accumulated data INST – instantaneous data MEAN – mean data In the future, this parameter
may be made optional if the time scale can be determined from the data type. None – must be specified. NewTimeScale The time scale for the new time series (see OldTimeScale for possible
values). In the future, this parameter may be made optional if the time scale can be determined from the data type. None – must be specified. Statistic Used in the case where INST (small
interval) to INST (large interval) conversion is occurring. A sample of values from the input time series, corresponding to the output interval, is determined and used to compute a statistic
instead of a simple value transfer. Statistics that are currently supported are MAX and MIN. The HandleMissingInputHow parameter is initially used to adjust missing data and then the
AllowMissingCount parameter is used to check whether the statistic can be computed. The statistic is determined from the old and new time scales. OutputYearType The output year type
if the output time series has an interval of Year. The output year type can only be specified for input time series having an interval of Calendar 93 Command Reference – ChangeInterval()
-13
ChangeInterval() Command TSTool Documentation Parameter Description Default Day or Month and the output can have a time scale of ACCM (sum the input values) or MEAN (average the input
values). The AllowMissingCount and AllowMissingConsecutive parameters are recognized. NewDataType The data type for the new time series. This will be set in the identifier of the new
time series. Use the data type from the original time series. NewUnits The units for the new time series. This will be set in the identifier of the new time series. Use the units from
the original time series. Tolerance Currently used when converting large interval MEAN data to small interval INST data. After the new time series is created, the volume of the new time
series over each old interval is compared to the old time series for that same interval. If the difference between the two is outside the specified tolerance percentage, then each value
in the new time series is adjusted so the totals will match. The endpoints are averaged for this comparison. Additionally, when the adjustment is made, the new starting value is averaged
with the ending value of the previous interval so that the previous interval is not overly affected by this calculation. 0.01 Handle EndpointsHow Indicates how endpoints should be handled
when changing from INST to MEAN, small interval to larger interval (daily output or finer), one of: AverageEndpoints – use both endpoint values for new single value IncludeFirstOnly
– only use earlier endpoint Average Endpoints AllowMissing Count The number of missing values allowed in the input interval in order to produce a result. For example, if converting daily
data to monthly, a value of 5 would allow <= 5 missing daily values and still compute the result. This capability should be used with care because it may result in data that are not
representative of actual conditions. This parameter is considered after the HandleMissingHow parameter. 0 – do not allow any missing data in the source data when computing a result.
AllowMissing Consecutive The number of consecutive missing values allowed in the input interval in order to produce a result. For example, if converting daily data to monthly, a value
of 3 would allow <= 3 consecutive missing daily values and still compute the result. The value must be less than or equal to AllowMissingCount. This parameter is considered after the
HandleMissingHow parameter. If not specified, the default for the number of allowed consecutive missing values is set to AllowMissingCount. OutputFill Method Use to fill output when
converting from INST to MEAN, large interval time series to small interval time series, one of: Repeat Command Reference – ChangeInterval() -14 94
TSTool Documentation ChangeInterval() Command Parameter Description Default Interpolate – linearly interpolate Repeat – repeat values for the output HandleMissing InputHow Indicate how
to handle missing values in input, one of: KeepMissing – leave data missing Repeat – repeat last non-missing value SetToZero – set values to 0 The missing data is handled on input and
the replacement value, if any, is applied to input and used for calculations just as if it was the actual value. The following cases do not use this parameter: • Irregular data • Day
and Month input converted to ACCM and MEAN. KeepMissing Several example command files follow. The following commands creates a Day ACCM time series from a Month ACCM time series: 0109.NOAA.Precip.Day
~HydroBase ChangeInterval(Alias=”0109Month”,TSList=AllMatchingTSID, TSID="0109.NOAA.Precip.Day",NewInterval=Month,OldTimeScale=ACCM,NewTimeScale=ACCM) The following commands create a