Precipitation data observed once-daily at 137 stations across New York were used to produce most products (Table 1). An additional 107 stations were utilized from surrounding states to facilitate interpolation during spatial mapping and for greater accuracy during formulation of common station subregions (Table 2). Nearly all stations were selected based on the following criteria:
1) At least 30 years of daily precipitation data;
2) Less than 10% missing data throughout both
a. the stationŐs entire period of record, and
b. the most recent 30 years.
These criteria were relaxed for about 5% of the stations included in the analysis in order to obtain necessary spatial coverage in a couple of sparse regions of the state. Also, 8 additional stations (not meeting the data completeness criteria) were used only to help determine subregional borders, but were not used when constructing return period maps, graphs or tables (Table 3).
Hourly precipitation data were necessary to produce some products. These data were used to determine weekly probabilities of exceeding extreme precipitation thresholds for shorter accumulation periods (6-hr, 3-hr, 2-hr, 1-hr). There is a lower density of stations that observe precipitation on an hourly basis compared to stations that only use once-daily observations. Of these available hourly stations, 42 were used in appropriate analyses, often with data available throughout the period 1953-2003 (Table 4).
Snowpack Water Equivalent (SWE) data were utilized in order to investigate the importance of melting (or accumulation) of snowpack on extreme precipitation return periods. SWE is measured once-daily at first-order stations, and 11 such stations in and surrounding New York State were used in these analyses (Table 5). SWE data at most of the stations were available starting in 1953.
Daily amounts of precipitation larger than 5 inches were tested for consistency with other nearby stations. Precipitation amounts of 10 inches or more were only considered valid if at least two stations within 200 miles observed at least 5 inches of precipitation on the same day, the previous day or the following day. For testing the validity of precipitation amounts between 5 inches and 10 inches, identical methods were used except the precipitation threshold at nearby locations was lowered to 3 inches. This method was similar to that used by Wilks and Cember (1993).
Quality control of SWE data utilized an automated procedure outlined by Schmidlin et al. (1995). This procedure checks for exceedence of specified extreme limits, recording errors, consistency with daily precipitation, estimated melt, possible drifting and expected measurement variability. Readers are referred to Schmidlin et al. (1995) for a detailed outline of the quality control procedure.