The longevity, completeness and spatial coverage of the database generated for this study is unique and unprecedented for New Jersey climate station records. Cooperative Station observational records are archived at the National Climatic Data Center (NCDC) in Asheville, NC. Prior to archiving, they are subjected to a rigorous quality control that includes looking for outliers and compares observations with those at surrounding stations. Suspect temperature data are flagged and replaced with a value based on inter-station comparisons. Suspect precipitation data are flagged and eliminated from the database, resulting in missing observations (which we have filled as part of our development process).
The ONJSC maintains a copy of the NCDC database in Structured Query Language (SQL) format. At the start of this phase of the project, only official data through August 2003 were available, thus 2003 is the end date of this evaluation. The ONJSC database was evaluated for the selection of primary and surrogate stations. The evaluation of candidate stations took into account the location and longevity of station records within the database. We strove to find stations that continue to observe today. Only Woodstown (5/31/2003) and Pemberton (2/28/2002) no longer functioned as Cooperative stations by August 2003. Unfortunately, since that time the Little Falls station has also closed. In some cases, a lack of extremely lengthy records for a region of the State necessitated the selection of stations with only just over 50 years of observations.
Once the 19 primary stations were selected (noting that only lengthy precipitation data are available at New Milford and Toms River), we conducted our own quality assessment, and were pleased to find that very few suspicious observations (which were discarded) were found. With none of the primary stations having 100% data completeness, the next step was to identify nearby surrogate stations. The experience of the principal investigator was employed here. Surrogates may not still be reporting, and may have closed years ago. However they were climatologically similar to the primary station, thus served as excellent substitutes when the need arose.
Surrogates were ranked according to their similarity to the primary station and the potential of filling in as much missing data as possible from one site. Software was developed to fill gaps, moving through the ranked surrogates until a primary station record was temporally complete. The temporally complete time series for these key NJ stations may represent the most important contribution of this study, hopefully serving many as they explore local and regional climate change and applied climate issues. This special database is available to all interested parties.
Time Series Development
The purpose of this project was to look at time series of temperature and precipitation beyond the "standard" approach of using annual means and totals. We attempted to be creative in selecting seasonal and threshold approaches to assess NJ climate variability and change. Limited study resources necessitated the selection of a subset of the myriad approaches that came to mind, however we believe a representative sampling of key variables has been completed (table2).
SQL routines were written to extract the base variable (e.g. daily temperature for a given station) from the database and calculate the selected study variable (e.g. heating degree days). Winter and heating degree day (HDD) evaluations overlap from one calendar year into another (e.g. the HDD season runs from July 1 of year 1 to June 30 of year 2; winter runs from December of year 1 to February of year 2). Spring includes the months of March-May, summer includes June-August, and fall runs from September-November.
Some 626 time series graphs were created to display project variables. They are available on the project web site by variable (all stations) or by individual station (all variables).
|return to Main Page|