Laserfiche WebLink
on the monthly rainfall totals, as that could provide indications of possible seasonal <br />variations. <br />A total of 11 stations in the NDCMP target areas and 25 stations in (10 of the) 12 <br />easternmost counties of Montana have remained in the same location and currently report <br />rainfall data as part of the climatic network. An additional gage at Marmarth, in western <br />Slope County, and two gages in Hettinger County, North Dakota, could not be used <br />because those areas have dropped out of the NDCMP in recent years. Unfortunately, the <br />period of record varies greatly; while records for a few of the stations go back to 1894, <br />others only began reporting in 1950 or even later. Moreover, one station in the target area <br />is lacking data for 15 years of the NDCMP era, while one in the (north) control area is <br />lacking 20 years of data from the 1950s to 1970. Table 1 illustrates the decrease in the <br />number of available stations with reasonably complete records as the period of interest is <br />extended further back in time. <br />Table 1: Availability of Climatic Rain Gage Data <br />Start Year Number of Stations with Usable Data <br />TargetControl(N/S) <br />19501021(10/11) <br />19491016(8/8) <br />19481011(6/5) <br />1946911(6/5) <br />1942810(5/5) <br />1938710(5/5) <br />193668(4/4) <br />To begin the analysis with an earlier start date would provide a longer period of <br />record with no seeding in the area (there has been some cloud seeding activity in western <br />North Dakota most years since 1951) but a smaller number of gages with the requisite <br />data continuity. After review of the information in Table 1, we decided to begin the <br />analysis with data from 1950. Trying to extend back to include even two more years <br />would lose almost half of the control-area stations. This choice of the analysis period <br />does involve a risk of diluting possible indications of any seeding effect, in view of the <br />seeding activities in the target region since 1951. However, there appeared to be no viable <br />alternative as the number of available gages even from 1950 is already small for <br />obtaining good estimates of the rainfall in the area. <br />The definition of the “target area” for this analysis also warrants some discussion. <br />About District 2 there is little question; all three counties (McKenzie, Mountrail, and <br />Ward) have participated continuously in the NDCMP since its inception. However, there <br />has been considerable variation in the makeup of District 1. Hettinger County was part of <br />the district only from 1976 through 1988, so was not included in this analysis because <br />that is barely half the NDCMP time period. Bowman County participated in all years <br />3 <br />