TSTool_User_09-09-00 - Laserfiche WebLink

TSTool Documentation 2.3 Time Scale for Time Series Data The time scale for time series data gives an indication of how the data value were measured or computed. The time scale is generally determined from the data type (or the data type and interval) and can be one of the following (the abbreviations are often used in software choices): Instantaneous (INST): The data value represents the data observed at the time associated with the reading (e.g., instantaneous temperature, streamflow, or snow depth). Instantaneous data may be of irregular or regular interval, depending on the data collection system. If irregular, the precision of the date /time associated with the reading may vary (e.g., automated collection systems may have very precise times whereas infrequently recorded field measurements may only be recorded to the nearest day). Accumulated (ACCM): The data value represents the accumulation of the observed data over the preceding interval. The date /time associated with the data value corresponds to the end of the interval. For example, precipitation (rain or snow recorded as melt) is often recorded as an accumulation over some interval. Accumulated values are typically available as a regular time series, although this is not a requirement (e.g., precipitation might be accumulated between times that a rain gage is read and emptied). Mean (MEAN): The data value represents the mean value of observations during the preceding interval. The date /time associated with the data value corresponds to the end of the interval. The mean includes values after the previous timestamp and including the current timestamp. The computation of mean values may be different depending on whether the original data are irregular or regular. For example, if the original data are regular interval, then equal weight may be given to each value when computing the mean (a simple mean). If the original data are irregular interval, then the weight given to each irregular value may depend on the amount of time that a value was observed (a time - weighted mean, not a simple mean). Without having specific information about the time scale for data, TSTool assumes that all data are instantaneous for displays. If time series are graphed using bars, an option is given to display the bar to the left, centered on, or to the right of the date /time. If time series are graphed using lines or points, the data values are drawn at the date /time corresponding to data values. This may not be appropriate for the time scale of the data. In most cases, this default is adequate for displays. Graphing data of different time scales together does result in visual inconsistencies. These issues are being evaluated and options may be implemented in future releases of the software. In particular, an effort to automatically determine the time scale from the data type and interval is being evaluated. This can be difficult given that data types are not consistent between input types and time scale may be difficult to determine when reading time series. Refer to the input type appendices for information about time scale. The time scale is particularly important when changing the time interval of data. For example, conversion of instantaneous data to mean involves an averaging process. Conversion of instantaneous data to accumulated data involves summing the original data. Commands that change interval either operate only on data of a certain time scale or require that the time scale be specified to control the conversion. Refer to the command documentation for specific requirements. 2.4 Time Series Commands and Processing Sequence Although TSTool can be run in batch mode (see Chapter 3 — Getting Started), you should be able to perform all time series viewing and manipulation within the GUI. Commands are used to read, manipulate, and output time series. Commands are processed sequentially from the first to the last Introduction - 6 26