The timestamp alignment module examines the alignment between the measured incoming radiation, e.g., photosynthetically active radiation (PPFD_IN), shortwave radiation (SW_IN), and the calculated potential incoming radiation at the top of the atmosphere (SW_IN_POT).
The module is used to identify the following issues:
- Wrong timestamp specification
- Misspecified beginning or ending timestamps
- Timestamps not matched with time zone specification
- Use of daylight saving time
- Data streams not synchronized
- Radiation measurement issue
- Sensor not leveled
- Shaded radiation measurements
- Higher than expected radiation readings
For each site, (half-)hourly SW_IN_POT for the entire data record is calculated based on a site’s geo-location and time zone (local standard time without daylight saving). The SW_IN_POT, SW_IN, and PPFD_IN data are aggregated into a “maximum diurnal composite” for each of the 15-day non-overlapping windows (1 and 2 in figure below). This maximum diurnal composite eliminates the influences of cloudy conditions to focus the alignment analysis on clear-sky conditions. The module expects that diurnal patterns align in time between SW_IN_POT, SW_IN, and PPFD_IN indicated by cross-correlation (4), or that at least SW_IN and PPFD_IN do not consistently exceed SW_IN_POT in the early-morning or late-afternoon hours (3).
Example figure illustrating the timestamp alignment check in which no issues are detected The “maximum diurnal composite” is calculated for each 15-day non-overlapping window within a year (1). The module expects that the diurnal patterns will align in time between SW_IN_POT and SW_IN as indicated by cross-correlation (3), or that at least SW_IN does not exceed SW_IN_POT in most periods (2).
Two types of statistics are calculated by the module to detect potential issues in timestamp alignments and radiation measurements:
- Percentages of occasions that SW_IN (or PPFD_IN) exceeded SW_IN_POT.
- Cross-correlations of the composite time series between SW_IN_POT and SW_IN (or PPFD_IN). For example, a maximum cross-correlation occurring at a two-step lag (i.e., one hour for data reported in 30-minute periods) suggests a one-hour shift in the timestamps, which could be attributed to a mismatch of time zone specifications.
Example figure illustrating the timestamp alignment module with detection of potential issues. This example shows (1) a number of occasions that observed radiation (SW_IN, PPFD_IN) exceeds the potential incoming radiation (SW_IN_POT), i.e., earlier sunrise/sunset than expected. And, (2) the maximum cross-correlation between calculated and observed radiation occurs at a two-step lag, i.e., a one-hour shift for 30-minute data.