Comparison of Quality and Completeness of National Climatic Data Center and Illinois Climate Assistance Service (CLASS) Data

Abstract

We compare the Completeness and quality of one year's daily temperature and precipitation observations from the National Climatic Data Center (NCDC) database with those received by the Illinois State Water Survey's (ISWS) Climate Assistance Service (CLASS). The NCDC data are keyed by hand from the E-15 forms received from the National Weather Service (NWS) Cooperative Observes, and are quality controlled. The CLASS data are transmitted by NWS Cooperative Observers via touch-tone phone to a Water Survey computer. These data are quality controlled only to the extent that each value is verbally repeated to the transmitting observer by voice synthesizer, and the ISWS compute checks to determine that the temperatures are consistent, i.e., maximum ≥ minimum and so on. The NCDC database was more complete than that of CLASS. This situation was most obvious for precipitation entries where CLASS observers typically made no precipitation entry (rather than transmitting a zero) to save transmission time. National Climatic Data Center, on the other hand, automatically enters zeros when precipitation was unreported but was unlikely to have occurred. The NCDC database which is rigidly quality controlled was also more accurate (only 0.4% errors) that class of CLASS (3.6% errors). It is possible that if telephone transmission became the official mechanism for data entry, missing data would be much reduced; and if each entry value had to be repeated by the observer, the accuracy of such transmissions would also improve.