Laserfiche WebLink
<br />estimates for that time period were likely accurate to within about 0.20 inches. While there <br />were only 69 locations which had actually measured the rainfall specific to the 1600 - 2000 <br />MDT time period July 27, confident estimates of rainfall for that period were made for an <br />additional 116 locations. Similarly, rainfall patterns Sunday night into Monday morning were <br />quite systematic, increasing nearly unifonnly from east to west across Fort Collins. As a <br />result, confident estimates could again be made. Table 5 shows the number of point estimates <br />made during each rainfa1l period and the total number of point values: estimates plus <br />measurements. <br /> <br />Table 5. <br />Number of point estimates of rainfall for each rainfall period during the July <br />27-28,1997 heavy rain episode in eastem Larimer County, Colorado. <br /> <br />Rainfall Period <br />1) Sunday evening <br />2) Overnight <br />3) Monday morning <br />Periods 1-3 combined <br /> <br />4) Monday evening <br />Total2-day Rainfall <br /> <br />Number of Confident <br />Point Estimates <br />116 <br />127 <br />98 <br />97 <br /> <br />85 <br />o <br /> <br />Estimates <br />+ Measurements <br />185 <br />169 <br />172 <br />199 <br /> <br />211 <br />251 <br /> <br />To check the validity of rainfall estimates for each measurement point, estimated rain amounts <br />for each of the rainfall periods were summed and compared to the measured precipitation <br />total. When the sum of the rainfa1l estimates for each rain episode did not match the total <br />observed rainfa1l at a particular location, then a more thorough site investigation was <br />performed. Data for each location within approximately one mile of the point in question <br />were examined. Based on the surrounding data, the confidence of each interval estimate was <br />evaluated. The most confident estimates were left unchanged while the least confident <br />estimates were modified so that the sum of each interval estimate would then equal the <br />observed storm total. Many minor adjustments to initial estimates were required to "force" <br />the sum of estimated values for the individual rain episodes to equal the observed rainfa1l total <br />for the entire storm. Most of these adjustments were less than 0.50 inches and did not <br />significantly alter the subsequent analyzed rainfall patterns. However, over the western <br />portion ofFort Collins and areas of eastern Larimer County close to the first hogback where <br />the eastern foothills begin, some adjustments of two inches or more were required to make <br />totals match. When these adjusted values were replotted and new rainfall patterns drawn, <br />significant differences were found. As an example, adjusted estimates suggested much heavier <br />rainfall totals during the Sunday night through Monday morning period in the area northwest <br />of Fort Collins near the town of Laporte. Reports of severe local flooding near Laporte <br />during the day on Monday, July 28, provided good confirmation that these adjustments were <br />appropriate. <br /> <br />Using both the observed data and estimates, a new set of maps was produced. Figure 5 shows <br />an example of a computer-generated analysis with and without estimated data. Confidence in <br /> <br />16 <br />