Due to the very good quality of the data gathered in New Jersey over the past century or longer, the number of primary stations selected for this study is approximately double the number projected in the study proposal. Temporally complete station records were generated for 19 stations (map). These primary stations were culled from a quality-controlled database that contains observations from over 100 NJ stations (many no longer operating), and were chosen due to their longevity and locations. The observation period ranged from 1893-2003 at 8 stations. The period was as short as 1948-2003 at the Atlantic City Marina station. These stations are part of the Cooperative Observing Station network of the National Weather Service.
Information on station location, elevation and start date is found in table 1. Also included are the percentage of the station record interval that is populated with primary station data, and the percentage of data that was needed from surrogate stations to make the primary record temporally complete. Surrogates were selected based on their proximity and climatic similarity to the primary station, and on the time interval(s) that required data filling. By being climatically similar there was no need to adjust surrogate data to match the climate of the primary location. Percentages were generated for temperature and precipitation observations. Primary observations populated from virtually 100% of the record at Newark (1 day missing) and 99.9% of the record at New Brunswick (1893-2003) to 86.5 % at Sussex (1893-2003). Pie diagrams depict percentages of primary and surrogate temperature and precipitation data contributions to the study station records. Some records required only one surrogate to attain completeness, four records required as many as four surrogates. Pie charts for each primary station depict the amount of filling needed to create temporally complete temperature and precipitation time series. Note that temperature and precipitation data may differ slightly in the need for filling.
When evaluating time series of station data, it is important to know as much as possible about the station itself. Such metadata include historical information on meteorological equipment, station moves and the time at which observations are taken each day. (Station Metadata) A change in any of these must be considered whenever a change in a station record is recognized, as they might be a contributing factor. Such "artificial" changes must be discounted prior to any climatological explanation regarding a change is rendered. Given the longevity of the station records employed in this study it is a given that changes have occurred at all stations.
For instance, many stations switched from late afternoon (e.g. 6PM) to early morning (e.g. 7AM) observations in the late 1960s. This results in a slight reduction in mean monthly minimum and maximum temperature. Occasionally a cold low temperature on a given day might appear twice in the record, first for the day ending at the time of observation and then for the next day (assuming tomorrow morning's low is not as low or lower than the temperature at the time the thermometer is reset on the first morning). This would not occur with a PM observation time, though in this case a potential "double high maximum" might be repeated, resulting in an inflated mean monthly maximum.
Many stations had liquid in glass thermometers replaced by electronic thermisters in the 1980s. Tests have shown that the latter record slightly lower maxima and higher minima than the liquid instruments.
All stations in this study have moved at least once. The most stable station has been New Brunswick (Newark is similar). The New Brunswick station only moved about a half mile (with little change in elevation) in the late 1960s and moved back to the original site in the late 1970s. Other stations, such as Flemington, have had more significant moves.Finally, a station may not move or not move too far, yet the local or regional surroundings of the station may change through time. This mainly manifests itself in elevated temperatures as a result of a heat island effect (assuming that change was associated with human development). Previous studies have suggested that Newark and Plainfield are the two study stations with such recognized urban-influenced warming. Such a change at all other locations has not been previously recognized, but must be considered in the evaluative process.
|return to Main Page|