/'\ 
 
 . - 
 
 NOAA Technical Memorandum NESS 81 
 
 ^ T0Fc Q 
 
 S ?AT ES O* 
 
 ESTIMATION OF DAILY PRECIPITATION 
 OVER CHINA AND THE USSR USING 
 SATELLITE IMAGERY 
 
 Washington, D. C. 
 September 1976 
 
 noaa 
 
 NATIONAL OCEANIC AND 
 ATMOSPHERIC ADMINISTRATION 
 
 / 
 
 National Environmental 
 Satellite Service 
 
NOAA TECHNICAL MEMORANDUMS 
 
 National Environmental Satellite Service Series 
 
 The National Environmental Satellite Service (NESS) is responsible for the establishment and oper- 
 ation of the environmental satellite systems of NOAA. 
 
 NOAA Technical Memorandums facilitate rapid distribution of material that may be preliminary in nature 
 and so may be published formally elsewhere at a later date. Publications 1 through 20 and 22 through 25 
 are in the earlier ESSA National Environmental Satellite Center Technical Memorandum (NESCTM) series. 
 The current NOAA Technical Memorandum NESS series includes 21, 26, and subsequent issuances. 
 
 Publications listed below are available from the National Technical Information Service, U.S. Depart- 
 ment of Commerce, Sills Bldg., 5285 Port Royal Road, Springfield, Va. 22151. Prices on request. Order 
 by accession number (given in parentheses) . Information on memorandums not listed below can be obtained 
 from Environmental Data Service (D831), 3300 Whitehaven St., NW., Washington, D.C. 20235. 
 
 NESS 41 Effect of Orbital Inclination and Spin Axis Attitude on Wind Estimates From Photographs by 
 Geosynchronous Satellites. Linwood F. Whitney, Jr., September 1972, 32 pp. (COM-72-11499) 
 
 NESS 42 Evaluation of a Technique for the Analysis and Forecasting of Tropical Cyclone Intensities From 
 Satellite Pictures. Carl 0. Erickson, September 1972, 28 pp. (C0M-72-11472) 
 
 NESS 43 Cloud Motions in Baroclinic Zones. Linwood F. Whitney, Jr., October 1972, 6 pp. (C0M-73- 
 10029) 
 
 NESS 44 Estimation of Average Daily Rainfall From Satellite Cloud Photographs. Walton A. Follansbee, 
 January 1973, 39 pp. (COM-73-10539) 
 
 NESS 45 A Technique for the Analysis and Forecasting of Tropical Cyclone Intensities From Satellite 
 Pictures (Revision of NESS 36). Vernon F. Dvorak, February 1973, 19 pp. (COM-73-10675) ' 
 
 NESS 46 Publications and Final Reports on Contracts and Grants, 1972. NESS, April 1973, 10 pp. 
 (COM-73-11035) 
 
 NESS 47 Stratospheric Photochemistry of Ozone and SST Pollution: An Introduction and Survey of Se- 
 lected Developments Since 1965. Martin S. Longmire, March 1973, 29 pp. (C0M-73-10786) 
 
 NESS 48 Review of Satellite Measurements of Albedo and Outgoing Long-Wave Radiation. Arnold Gruber, 
 July 1973, 12 pp. (COM-73-11443) 
 
 NESS 49 Operational Processing of Solar Proton Monitor Data. Louis Rubin, Henry L. Phillips, and 
 Stanley R. Brown, August 1973, 17 pp. (COM-73-11647/AS) 
 
 NESS 50 An Examination of Tropical Cloud Clusters Using Simultaneously Observed Brightness and High 
 Resolution Infrared Data From Satellites. Arnold Gruber, September 1973, 22 pp. (COM-73- 
 11941/4AS) 
 
 NESS 51 SKYLAB Earth Resources Experiment Package Experiments in Oceanography and Marine Science. A. 
 L. Grabham and John W. Sherman, III, September 1973, 72 pp. (COM 74-11740/AS) 
 
 NESS 52 Operational Products From ITOS Scanning Radiometer Data. Edward F. Conlan, October 1973, 57 
 pp. (COM-74-10040) 
 
 NESS 53 Catalog of Operational Satellite Products. Eugene R. Hoppe and Abraham L. Ruiz (Editors), 
 March 1974, 91 pp. (COM-74-11339/AS) 
 
 NESS 54 A Method of Converting the SMS/GOES WEFAX Frequency (1691 MHz) to the Existing APT/WEFAX Fre- 
 quency (137 MHz). John J. Nagle, April 1974, 18 pp. (C0M-74-11294/AS) 
 
 NESS 55 Publications and Final Reports on Contracts and Grants, 1973. NESS, April 1974, 8 pp. 
 (COM-74-11108/AS) 
 
 NESS 56 What Are You Looking at When You Say This Area Is a Suspect Area for Severe Weather? Arthur H. 
 Smith, Jr., February 1974, 15 pp. (COM- 74-11333/ AS) 
 
 NESS 57 Nimbus-5 Sounder Data Processing System, Part I: Measurement Characteristics and Data Reduc- 
 tion Procedures. W.L. Smith, H. M. Woolf, P. G. Abel, C. M. Hayden, M. Chalfant, and N. Grody, 
 June 1974, 99 pp. (COM- 74-11436/ AS) S 
 
 NESS 58 The Role of Satellites in Snow and Ice Measurements. Donald R. Wiesnet, August 1974, 12 pp. 
 (C0M-74-11747/AS) 
 
 (Continued on inside back cover) 
 
NOAA Technical Memorandum NESS 81 
 
 ESTIMATION OF DAILY PRECIPITATION 
 OVER CHINA AND THE USSR USING 
 SATELLITE IMAGERY 
 
 Walton A. Follansbee 
 
 Washington, D. C. 
 September 1976 
 
 UNITED STATES 
 DEPARTMENT OF COMMERCE 
 Elliot L. Richardson, Secretary 
 
 NATIONAL OCEANIC AND 
 ATMOSPHERIC ADMINISTRATION 
 Robert M White, Administrator 
 
 National Environmental 
 
 Satellite Service 
 
 David S Johnson, Director 
 
Digitized by the Internet Archive 
 
 in 2012 with funding from 
 
 LYRASIS Members and Sloan Foundation 
 
 http://archive.org/details/estimationofdailOOfoll 
 
CONTENTS 
 
 Abstract 1 
 
 Introduction 2 
 
 Assumptions 2 
 
 Cloud motion models 2 
 
 Development of the technique 12 
 
 Procedure 13 
 
 Tests and verification 16 
 
 Future investigations 24 
 
 Acknowledgments 29 
 
 References 30 
 
 in 
 
ESTIMATION OF DAILY PRECIPITATION OVER CHINA 
 AND THE USSR USING SATELLITE IMAGERt 
 
 Walton A. Follansbee 
 National Environmental Satellite Service, NOAA, Washington, D.C. 
 
 ABSTRACT. This memorandum describes a technique for 
 estimating precipitation over China and the USSR in 
 late autumn, winter, and early spring, using satel- 
 lite data. The data consisted of NOAA-4 infrared 
 imagery taken at approximately 9 a.m. and 8 p.m., 
 local time, and visual pictures taken near 9 a.m. 
 local time, all with one-half mile resolution. The 
 major human input to the technique is identification 
 of precipitating clouds in the infrared imagery, 
 aided by inspection of the visual pictures. 
 
 The technique consists of three steps: (1) delineate 
 the area(s) covered by precipitating clouds in each 
 of two consecutive infrared images taken at approxi- 
 mately 12-hour intervals; (2) using cloud motion 
 models developed for the purpose, delineate the 
 envelope of precipitation and isopleths of duration 
 of precipitation in hours; (3) using an empirically 
 derived relationship between duration of precipita- 
 tion and the percentage of monthly normal precipita- 
 tion, estimate amounts of precipitation for the 12- 
 hour period at each grid point in the area of inter- 
 est. The relationship is P 12 = 0.0075 DP N , where D 
 is duration in hours, Pj^ is monthly normal precipita- 
 tion, and Px2 is total precipitation in the 12-hour 
 period. 
 
 Satellite data for the period November 21, 1974, to 
 February 28, 1975, were used to develop the technique 
 and to establish the relationships. The technique 
 was then tested in a quasi-operational mode for the 
 months of March and April, 1975. For March 1975, the 
 ratio of total error to total observed precipitation 
 for the USSR was 0.55, and the ratio of algebraic 
 error to total observed precipitation was 0.01. This 
 degree of accuracy and the absence of significant 
 bias was repeated in the April estimates for the 
 USSR. A similar test over 24 States in the central 
 United States for April 1976 gave ratios of 0.46 and 
 -0.28. This underestimate is due largely to frequent 
 thunderstorms over the U.S. in April 1976. 
 
INTRODUCTION 
 
 The problem of estimating precipitation using satellite data has received 
 increasing attention by a number of investigators in recent years. To date 
 the emphasis has been on convective rain over the tropics and in summer in 
 the subtropics, usually for mesoscale areas. 
 
 The technique described herein was prompted by the need for estimting pre- 
 cipitation over vast agricultural regions of Asia in autumn, winter, and 
 spring. Convective rain is far less prevalent in these Asiatic areas during 
 the colder months than in regions investigated in previous studies. This 
 has both advantages and disadvantages. Except in well-defined frontal sys- 
 tems, nimbostratus is more difficult to identify than cumulonimbus. On the 
 other hand, precipitation from nimbostratus is far more uniformly distributed 
 than that from cumulonimbus. 
 
 ASSUMPTIONS 
 
 An examination of climatological records led to the assumption that, in a 
 normal month over the middle latitudes of Asia, two storms per week or nine 
 storms per month could be expected. If the storms were of uniform intensity, 
 each might contribute 11% of the monthly normal precipitation. Assuming 
 that 11% fell in a 24-hour day, about 6% could be expected in a 12-hour 
 period, provided the precipitation was continuous throughout the 12 hours. 
 An analysis of the data showed that for a given point the duration of precipi- 
 tation is usually less than 24 hours, suggesting that normally more than 6% 
 would occur in a continuous 12-hour rain or snow period. 
 
 In the interest of simplicity, the precipitation rate in all parts of a 
 rain cloud has been considered constant in most cases, especially during the 
 development of the technique. On the other hand, it is assumed that an 
 experienced satellite meteorologist will have the capability of determining, 
 in most cases, which clouds or parts of clouds are in fact precipitating, and 
 at times, of assigning different intensities to various parts of the cloud. 
 
 CLOUD MOTION MODELS 
 
 While it is usually not difficult to identify a precipitating cloud mass 
 in succeeding 12-hour pictures, the duration of precipitation along its path 
 during the 12 hours presents problems. Motion models were developed to cope 
 with these problems. 
 
 One of the simplest of such models would be a rectangular cloud with major 
 axis oriented north-south, moving from west to east at uniform speed, and 
 traveling in 12 hours a distance equal to its minor axis, without changing 
 shape or size. That is, at the end of 12 hours its trailing edge will reach 
 a line occupied by the leading edge at the beginning of the 12-hour period, 
 as shown in figure 1. Let us assume that the cloud is 100 miles long and 
 60 miles wide, and the satellite pictures have been taken at midnight and 
 noon on a given day. Then at 2 a.m. the trailing edge of the cloud will have 
 
MEAN DURATION OF PRECIPITATION (HOURS): 
 
 1 
 
 3 
 
 5 
 
 7 
 
 9 
 
 AT NOON 
 
 11 
 
 \- 
 x 
 
 2 
 
 Q 
 
 2 
 
 9 
 
 7 
 
 5 
 
 3 
 
 1 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 
 h- 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 
 <^ 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 CM 
 
 "* 
 
 to 
 
 03 
 
 o 
 
 C3 
 
 
 CM 
 
 ■* 
 
 (fl 
 
 CO 
 
 o 
 
 O 
 
 o 
 
 o 
 
 o 
 
 <— 
 
 Q 
 
 LU 
 
 o 
 
 o 
 
 o 
 
 o 
 
 •— 
 
 I- 
 
 h- 
 
 I- 
 
 H 
 
 H 
 
 LU 
 
 o 
 
 H 
 
 I- 
 
 H 
 
 h- 
 
 H 
 
 < 
 
 < 
 
 < 
 
 < 
 
 < 
 
 C3 
 
 2 
 
 < 
 EC 
 
 s 
 
 a 
 
 LU 
 
 a 
 
 2 
 
 a 
 < 
 
 LU 
 
 —I 
 
 < 
 
 < 
 
 < 
 
 < 
 
 < 
 
 ss 
 
 
 
 
 
 y 
 
 Ss> 
 
 
 
 
 
 
 ._^v*. 
 
 J 
 
 CLOUD BAND AT MIDNIGHT 
 
 CLOUD BAND AT NOON 
 
 Figure 1. --Cloud motion model, case 1. Rectangular, no overlap or underlap 
 
 moved 10 miles east of its midnight position, and a strip 10 miles by 100 
 miles will have cleared. Precipitation will have ended at midnight along 
 the western edge of this strip but will end at 2 a.m. along the eastern edge; 
 therefore the average duration of precipitation within the entire strip will 
 be 1 hour. But during this period all of the area extending 50 miles east of 
 the strip will have received precipitation throughout the 2 hours. In the 
 same way it can be shown that between 2 and 4 a.m. the 10-mile wide strip 
 immediately east of this first strip will have an average duration of 1 hour, 
 for a total of 3 hours duration from midnight to 4 a.m. Likewise, the next 
 strip to the east will have an average duration of 5 hours, and the next three 
 strips will have 7, 9, and 11 hours, respectively. 
 
 In like manner, following the progress of the leading edge of the cloud 
 mass, we see there would be six more strips 10 miles wide with average dura- 
 tion values in hours of 11, 9, 7, 5, 3, and 1, decreasing eastward. 
 
 A second case, shown in figure 2, is similar to the first, but traveling 
 in 12 hours a distance equal to twice its minor axis. At this speed, it can 
 be shown that duration of precipitation will not exceed 6 hours at any point 
 

 
 
 
 MEAN DURATION OF PRECIPITATION (HOURS): 
 
 
 1 
 
 
 3 
 
 5 
 
 6 
 
 6 
 
 6 
 
 5 
 
 h- 
 
 
 
 
 
 1- 
 
 
 
 
 I 
 
 
 
 
 
 X 
 
 
 
 
 U 
 
 
 
 
 
 o 
 
 
 2 
 
 o 
 
 
 2 
 
 
 
 
 
 2 
 
 
 o 
 
 
 Q 
 
 
 
 
 
 a 
 
 
 2 
 
 
 ^ 
 
 
 
 
 
 § 
 
 o 
 
 o 
 
 O 1- 
 o < 
 
 o 
 
 o 
 
 \- 
 
 
 o 
 
 o 
 
 o 
 
 H O 
 
 H o 
 
 *T ^ 
 
 <o 
 
 <r 
 
 
 o 
 
 o 
 
 o 
 
 < ° 
 
 O o 
 
 O UJ 
 
 o 
 
 
 
 CM 
 O 
 
 1- 
 
 o 
 
 (0 
 
 o 
 H 
 
 ^ CO 
 UJ o 
 
 2 H 
 
 1- 2 
 
 < Q 
 
 UJ 
 
 h- 
 < 
 
 UJ 
 
 
 < 
 
 < 
 
 < 
 
 Q < 
 
 UJ 
 
 < 
 
 o 
 
 
 a 
 
 
 
 
 
 C3 
 
 
 2 
 
 
 2 
 
 
 
 
 
 2 
 
 
 _i 
 
 
 _i 
 
 
 
 
 
 Q 
 
 
 < 
 
 
 < 
 
 1- 
 
 
 
 
 
 < 
 UJ 
 
 -J 
 
 
 CC 
 
 
 
 s!S 
 
 
 
 
 
 -s^ 
 
 
 
 
 
 
 \ 
 
 
 
 />> 
 
 
 
 
 
 ^* 
 
 ■>l 
 
 o 
 o 
 
 2 
 < 
 
 UJ 
 
 a 
 a 
 
 LU 
 (J 
 
 2 
 
 5 
 
 < 
 
 J 
 
 CLOUD BAND AT MIDNIGHT 
 
 CLOUD BAND AT NOON 
 
 Figure 2. --Case 2. Rectangular, under lap equal to width of band, 
 
 along the cloud's track. This case has an underlap* equal to the width of 
 the cloud band, or the cloud band's minor axis. 
 
 Figure 3 presents a third case of a rectangular cloud, moving in the direc- 
 tion of its major axis a distance of half the length of the major axis in 
 12 hours. This case provides a 50% overlap of the original cloud position 
 and the cloud position 12 hours later. If the major axis is 120 miles and 
 the minor axis 100 miles, the entire envelope of precipitation will be the 
 perimeter of a rectangle 100 by 180 miles. If, as before, the cloud is 
 moving from west to east, we could divide the entire precipitation area into 
 18 rectangular strips 10 miles wide and 100 miles long, such that the average 
 duration of precipitation in hours, beginning with the westernmost and ending 
 with the easternmost strip, would be 1, 3, 5, 7, 9, 11, 12, 12, 12, 12, 12, 
 12, 11, 9, 7, 5, 3, and 1. 
 
 Figure 4 shows cases similar to cases 1, 2, and 3, but for circular cloud 
 masses rather than rectangular. 
 
 *Underlap=the distance between the leading edge of the cloud mass at 
 initial time and the trailing edge of the same cloud mass at the end of 
 the time period under consideration. 
 
MEAN DURATION OF PRECIPITATION (HOURS): 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 1 
 
 1 
 
 3 
 
 5 
 
 7 
 
 9 
 
 11 12 
 
 2 
 
 o 
 o I 
 
 2 1 
 1- 1 
 
 12 
 
 12 
 
 12 
 
 12 
 
 12 
 
 11 
 
 I- 
 I 
 
 a 
 
 2 
 Q 
 
 2 
 
 9 
 
 7 
 
 5 
 
 3 
 
 1 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 < I 
 
 
 
 
 
 
 I- 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o 
 
 
 
 
 
 
 
 
 o 
 
 o 
 
 o 
 
 o 
 
 o I 
 
 CM 
 
 »* 
 
 9 
 
 CO 
 
 o 
 
 
 
 
 
 
 
 
 CM 
 
 ^- 
 
 (0 
 
 CO 
 
 o 
 
 o 
 
 © 
 
 o 
 
 o 
 
 
 
 
 
 
 
 
 LU 
 
 O 
 
 o 
 
 o 
 
 o 
 
 *— 
 
 I- 
 
 h- 
 
 \- 
 
 t- 
 
 I- 
 
 
 
 
 
 
 
 o 
 
 H 
 
 I- 
 
 K 
 
 H 
 
 H I 
 
 < 
 
 < 
 
 < 
 
 < 
 
 < 
 
 2 
 _l 1 
 
 < 1 
 
 GC 1 
 1- J 
 
 
 
 
 
 
 Q 
 HI 
 
 2 
 
 a 
 < 
 hi 
 
 < 
 
 < 
 
 < 
 
 < 
 
 < 
 
 *SS 
 
 
 
 
 
 \ 
 
 
 
 
 
 
 »" 
 
 
 
 
 
 — >l 
 
 J 
 
 ?S 
 
 
 
 
 
 1 
 
 zir 
 
 CLOUD BAND AT MIDNIGHT 
 
 CLOUD BAND AT NOON 
 
 Figure 3. --Case 3. Rectangular, overlap equal to 50% of band width. 
 
 Obviously, not many cloud masses retain a rectangular or circular shape 
 of uniform size for 12 consecutive hours. It is possible, however, to 
 adapt cases 1 through 6, singly or in combination, to many cloud masses as 
 they appear in satellite imagery. Figure 5 shows a somewhat idealized 
 frontal system moving uniformly with little change in shape or size. The 
 upper third is fairly circular and shows approximately 50% overlap; therefore 
 it may be treated like case 6. The middle third is fairly rectangular, with 
 little overlap or underlap, and may be likened to case 1. The lower third 
 remains nearly rectangular and averages an underlap approximately equal to 
 its width; thus it may be treated like case 2. The isopleths of precipita- 
 tion duration have been smoothed in regions of transition from one case to 
 the next . 
 
 Figure 6 shows examples of three cloud systems, cases 8, 9, and 10, which 
 either expand or contract without any other lateral motion. The concentric 
 isopleths of precipitation duration are drawn the same way regardless of 
 whether the system is contracting or expanding. 
 
 Figure 7a shows a slightly more complex but more realistic motion pattern, 
 case 11. The cloud band is diminishing as it moves along a diagonal. 
 Positions are shown at 3-hour intervals beginning at 0000, ending at 1200 
 hours. Beginning and ending times of precipitation are indicated for various 
 points and lines. Based on these and additional beginnings and endings (not 
 shown), isopleths of duration of precipitation are outlined in figure 7b. 
 
 The underlap in case 11 is approximately equal to the width of the cloud 
 band at initial time. Case 12, figure 8a, is similar to case 11 except that 
 there is p small overlap equal to one-third of the cloud band width at final 
 
5 5 5 5 5 5 
 
 TRAILING 
 
 EDGE 
 
 AT MIDNIGHT 
 
 LEADING 
 
 EDGE 
 AT NOON 
 
 5 5 5 5 5 5 5 
 
 TRAILING 
 
 EDGE 
 
 AT MIDNIGHT 
 
 LEADING 
 
 EDGE 
 AT NOON 
 
 TRAILING 
 
 EDGE 
 
 AT MIDNIGHT 
 
 LEADING 
 
 EDGE 
 AT NOON 
 
 Figure 4. --Circular cloud motion models. Top: Case 4, no overlap or under-- 
 lap at line of centers. Center: Case 5, underlap at line of centers equal 
 to diameter. Bottom: Case 6, 50% overlap along line of centers. 
 
CASE 6 
 
 CASE 2 
 
 Figure 5. --Case 7. 
 
 Composite cloud motion model, combining cases 
 1 , 2 , and 6. 
 
(a) 
 
 (b) 
 
 (c) 
 
 Figure 6. --Cloud motion models for systems which either expand 
 or contract without any other lateral motion. (a) Case 8, 
 circular; (b) Case 10, vortical; (c) Case 9, oval. 
 
OHRS 
 
 Figure 7a. --Case 11. Oval, 
 contracting and moving on a 
 diagonal to major axis, with 
 under lap approximately equal to 
 width of cloud band at initial time. 
 
 B12 
 
 B12 
 
 Figure 7b. --Case 11. Duration of 
 precipitation analysis. 
 
10 
 
 EOO 
 
 Figure 8a. --Case 12. Oval, contracting and moving on a diagonal to major 
 axis, with overlap equal to one-third of cloud band width at terminal 
 time. 
 
 Figure 8b. --Case 12. Duration of precipitation analysis. 
 
1 1 
 
 time. Isopleths of duration of precipitation, based on beginning and ending 
 times from figure 8a, are outlined on figure 8b. 
 
 Figure 9 is an actual case over mainland China on March 14, 1975. The 
 heavy solid line in figure 9a is the cloud outline at 9 a.m. local time; the 
 heavy dashed line is the cloud outline at approximately 8 p.m. local time. 
 (Because of orbital factors, there is not exactly a 12-hour interval between 
 northbound morning and southbound evening local picture-taking time. How- 
 ever, for this study the interval has been assumed to be 12 hours. Thus for 
 figure 9b, initial time is given as 0000, and terminal time as 1200 hours.) 
 
 ( 
 
 J 
 
 v / 
 
 1 
 
 8 P.M. 
 
 y 
 
 f 
 
 V. 
 
 / 
 
 -J 
 
 Figure 9a. --Cloud positions over China at 9 a.m. and 8 p.m. local time, 
 
 March 14, 1975. 
 
 B00E03 B03E06 
 
 \ B09E1; 
 
 B06E09 
 •••o-oo 
 
 Jfo°° o ^^o 
 
 .^^oo° o° 
 o 
 o 
 
 B03E09 (JB09E12 o° 
 
 »+t^^N °°°o 
 
 ' BD6EVU o0 B 2 9 o E ooo°° 
 
 / o° 
 
 X_- y nOOOO° 000000 ! — -" 
 
 B03E06 
 
 B"l2__J 
 
 r 
 
 A /" 0< > - q ° ■■-./:. \ -^ 
 
 5 B06E09 B09E12 
 
 Figure 9b. --Cloud positions at 3-hour intervals (from Fig. 9a). 
 
 In figure 9b the cloud position at 0300 hours is outlined in a light solid 
 line; at 0600 hours, in a light dashed line; and at 0900 hours, in a line of 
 small circles. These intermediate positions have been "eyeballed," taking 
 
12 
 
 Figure 9c. --Duration of precipitation analysis (from Fig. 9b) 
 
 into account a gradual evolution of the salient features of the cloud mass. 
 The envelope of motion is a dotted line. As in figure 7, beginning and 
 ending times for various points and lines are indicated. Isopleths of 
 duration of precipitation are shown in figure 9c. 
 
 DEVELOPMENT OF THE TECHNIQUE 
 
 The original assumption, that a point remaining under a precipitating 
 cloud for 12 hours should produce 6% of the monthly normal for that point, 
 was tested over the USSR for late November and early December 1974. When 
 this resulted in general underestimates, the figure was raised to 12%. This 
 produced overestimates of comparable magnitude. The figure ultimately 
 adopted was 9%. All subsequent tests, covering the period from November 24, 
 1974, to April 30, 1975, show little or no bias using 9% of the monthly 
 normal for precipitation durations of twelve hours. 
 
 The assumption that all portions of a precipitating cloud mass produce 
 equal precipitation intensities is, of course, an oversimplification. For 
 example, in bright (cold) cloud masses seen in infrared imagery taken over 
 middle latitudes, little or no surface precipitation occurs under that part 
 of the cloud shield lying east (downwind) of the upper level ridge line, 
 but considerable precipitation occurs to the west (upstream) of this line. 
 Therefore, whenever this ridge line could be located with reasonable accu- 
 racy, it was drawn as the eastern or downwind terminus of precipitating cloud. 
 
 One frontal system, obviously deepening rapidly, was assigned much larger 
 percentage values in its more active portions, with percentage assignments 
 tapering off toward the storm periphery. 
 
 In general, precipitation was considered less intense near the lateral 
 edges of the cloud mass--for example, near the northern and southern edges 
 of a system moving eastward. It seemed unwise, however, in the early stages 
 of the investigation and development, to introduce too much subjective judg- 
 ment. Therefore, with the above-mentioned exceptions, the rule adopted was 
 to assign equal intensities to all parts of a precipitating cloud. 
 
13 
 
 Figure 10. --Infrared imagery taken at approximately 9 a.m., 
 local time, March 22, 1975, over Asia. 
 
 Figure 11 . --Infrared imagery taken at approximately 8 p.m., 
 local time, March 22, 1975, over Asia. 
 
14 
 
 60N 
 
 55N 
 
 50N 
 
 45 N 
 
 CI 
 
 rsi 
 
 CM 
 O 
 
 u 
 
 Oh 
 CO 
 
 o 
 +-> 
 
 e 
 
 Oi 
 
 -a 
 o 
 
 •H 
 f-> 
 (D 
 
 Oh 
 
 <D 
 
 4-J 
 
 u 
 o 
 
 c- 
 
 CO 
 
 Oh 
 
 03 
 
 U 
 O 
 
 CM 
 
 •H 
 
15 
 
 PROCEDURE 
 
 An understanding of the procedure finally adopted is facilitated by refer- 
 ence to figure 12, the workmap for estimating precipitation over the USSR 
 for the period, 9 a.m. to 8 p.m. local time, March 22, 1975. 
 
 The outlines of precipitating clouds at 9 a.m., seen in figure 10, are 
 drawn on the workmap in any selected color (heavy solid lines in figure 12) . 
 Similarly, precipitating clouds at 8 p.m., seen in figure 11, are drawn on 
 the workmap in a different color (heavy dashed lines in figure 12). The 
 envelopes of cloud motion are drawn (dotted lines in figure 12). Areas of 
 overlap of precipitating clouds are shaded, and marked "9" to indicate that 
 in these areas 9% of the monthly normal precipitation is occurring during 
 the 12-hour period. 
 
 Lines are drawn bisecting the zones from the edges of the overlap areas to 
 the edges of the corresponding envelopes. The light solid lines, such as 
 that crossing 40°N, 40°E, are such bisecting lines. In zones bounded by the 
 bisecting line and the envelope, precipitation will be less than 4.5% of the 
 monthly normal, while in zones bounded by the bisecting line and the edge of 
 the overlap, precipitation will be greater than 4.5% but less than 9% of the 
 monthly normal . 
 
 All of these smaller zones are bisected in turn, and marked 1-2, 3-4, 5-6, 
 and 7-8, as appropriate, to indicate sub-zones receiving 1 to 2% of the 
 monthly normal, 3 to 4%, 5 to 6%, and 7 to 8%. The light dashed line cross- 
 ing 50°N, 75°E is an example of lines separating sub-zones receiving 1 to 2% 
 from those receiving 3 to 4% of the monthly normal. The light dashed line 
 through 50°N, 77°E separates sub- zones having 5 to 6% from sub- zones having 
 
 7 to 8%. When desirable, these sub-zones may be bisected to delineate areas 
 each averaging 1, 2, 3, 4, 5, 6, 7, or 8% of the monthly normal, but usually 
 this is unnecessary. 
 
 The assignment of exact integers to these zones is somewhat arbitrary, 
 since it can be shown that the exact means for zones designated 1., 2, 3, 4, 
 5, 6, 7, and 8% are, respectively, 0.56, 1.69, 2.81, 3.94, 5.06, 6.19, 7.31, 
 and 8.44%. The absolute mean "error" is 0.25%, but the algebraic "error" is 
 zero. The advantages of using whole numbers and exact bisections of the 
 zones far outweigh any advantages resulting from a more rigorous but unwieldy 
 treatment. No real error can be determined; the relationship used herein, 
 between duration of precipitation and percent of monthly normal precipitation, 
 is entirely empirical, hence arbitrary. 
 
 In some situations no overlap occurs during the 12-hour period. In others, 
 the weak systems observed must be assigned less intense rainfall rates. At 
 times an old system moves out of an area and is replaced by a new system 
 moving into the vacated area 12 hours later. Each of these cases is exempli- 
 fied in figure 13, which shows cloud motions over mainland China for the 
 period 9 a.m. to 8 p.m., March 10, 1975. Figures 14 and 15 are the infrared 
 imagery for this period. 
 
 In figure 13, the system moving off the coast between 26° and 37°N at 
 
 8 p.m. (the heavy dashed line) has been formed by the merging of 
 
16 
 
 three smaller systems farther west at 9 a.m. (heavy solid lines). Advection 
 is sufficiently rapid that none of these smaller clouds overlaps the 8 p.m. 
 boundary of the merged system. The small cloud near 31°N between 113° and 
 119°E almost overlaps; therefore an isopleth enclosing a zone of 7% of the 
 monthly normal is drawn from the nose of the small morning cloud well into 
 the larger evening cloud. Zones ranging from 6% down to 1% encircle the 7% 
 zone, out to the cloud motion envelope. The small cloud located south of 
 30°N at 9 a.m. fails to overlap by an even larger distance; therefore the 
 wettest zone connecting it to the 8 p.m. cloud mass shows a mere 5% of 
 monthly normal. Note how the isopleths are drawn between the two areas of 
 merging. The treatment is somewhat subjective, but is considered logical. 
 
 The small cloud system along 35°N from 98° to 112°E at 9 a.m. (in figure 13) 
 presents three special problems: (1) it moves rapidly and thus shows even 
 greater underlap than its two companions; (2) it is weak and must be assigned 
 less intense precipitation; and (3) as it moves rapidly eastward it is re- 
 placed by a new larger system whose nose reaches as far east as 32°N, 115°E 
 by 8 p.m. Had this new system not moved in, the small band along 35°N at 
 9 a.m. would have shown 1% of the monthly normal precipitation throughout its 
 length, except in its eastern extremity where a 2% zone connects it with the 
 8 p.m. position farther east. The increasing strength of the cloud with time 
 makes it advisable to draw a small 4% zone along 35°N between 116° and 118°E. 
 Meanwhile the new large system moving in from west of 100°E is drawn with 
 intensities ranging from 1% to 5%. However, when the new system moves into 
 areas previously occupied by the small weak system, graphical addition is 
 used to show larger percentages, or greater total rainfall. 
 
 The system in extreme northeastern China in figure 13 is weakening with 
 time. Thus the overlap area is restricted to 1% of the monthly normal 
 rather than the usual 9%. 
 
 When percentages have been assigned to all precipitation areas, the per- 
 centage value at each grid point is multiplied by that grid point's monthly 
 normal precipitation to obtain the estimated 12-hour precipitation. Or, if 
 the weekly precipitation estimate is required for each grid point, percentages 
 at the grid points may be entered on data sheets such as that shown in fig- 
 ure 16. Percentages for each of the fourteen consecutive 12-hour periods of 
 the week are entered on the fourteen numbered lines. Grid point numbers 
 appear at the head of each column. The total for each column is multiplied 
 by the grid point normal for the month to get the precipitation estimate for 
 the week. Alternatively, once the map has been completed, the grid point 
 data may be scanned and stored by computer for subsequent retrieval. 
 
 TESTS AND VERIFICATION 
 
 The method was tested in a quasi-operational mode for the USSR and mainland 
 China for the period March 1 to April 30, 1975. The author made twice-daily 
 estimates of precipitation for 660 Russian and 450 Chinese grid points, 
 summed these to get weekly precipitation estimates at each grid point, and 
 drew isohyetal analyses of the estimates for each week of the period. The 
 estimates for March 1975 were begun on March 4, 1975 (when the first imagery 
 
17 
 
 <#*<f* 
 
 100E 
 
 A10E 
 
 Figure 13. --Workmap, China, for the period 9 a.m. to 8 p.m., local time, 
 
 March 10, 1975 
 
IS 
 
 x « 
 
 O LO 
 
 P-c-t 
 aj 
 
 +-> o 
 aj t— i 
 
 C X 
 0) o 
 
 -^ u 
 
 03 aj 
 +-> S 
 
 X - 
 
 U <D 
 O B 
 bO-H 
 aj 4-> 
 
 e 
 
 0) 
 Jh 
 
 aj 
 
 c 
 
 Ph • 
 
 cj 
 
 00 -H 
 
 t/) 
 
 X< 
 
 fn +-> 
 
 3 aj 
 
 bO £ 
 •H 
 
 O LO 
 
 QhOI 
 0_,<-l 
 
 aj 
 
 4-> O 
 Oj r— I 
 
 C X 
 CD O 
 
 aj aj 
 
 X - 
 
 o g 
 
 bO-H 
 aj +j 
 £ 
 
 <D 
 U 
 
 aj 
 in 
 
 P! 
 
 i aj . 
 i aj 
 
 • en -h 
 
 0) CD 
 
 b t? 
 
 3 aj 
 W> £ 
 ■H 
 
 Ph 
 
19 
 
 LO 
 
 
 4-> 
 
 oo 
 
 c 
 
 
 •H 
 
 
 o 
 
 
 a, 
 
 
 T3 
 
 r^ 
 
 ■H 
 
 
 ^ 
 
 
 O 
 
 
 vD 
 
 to 
 
 o o <u 
 
 • fn g 
 
 <n a> 3 
 
 to Ol «tf 
 
 r^a^rH cm en r^ t-i v£> 
 
 LO o 
 
 ^ o 
 
 n- en cm rsj^-t^r^i— t <3- oo *3- cm oo 
 
 lo o 
 
 LO oo 
 
 o lo 
 
 ■rj-r^ooaii— i logics oo en 
 
 LO o 
 
 v0 
 
 lo oo oo r^ 
 
 N O^ Ol (N 
 
 \0 to 
 
 oo t-» 
 
 ^ o 
 
 r-H LO 
 
 00 CT> CM <— I O CM 
 
 CM ,— I 
 
 O 
 
 00 o 
 I— I 
 
 o 
 
 to 
 
 o 
 
 o 
 
 LO 
 
 o 
 
 to 
 
 o 
 
 o 
 
 CN 
 
 o 
 
 to 00 
 
 i— I LO L0 
 
 oo 
 
 CM 00 a> 
 
 -3- vo lo 
 
 ■-1 o 
 
 \0 CM O 
 
 tO r-H 
 
 •>3" Cn Cn Cn 
 
 to O 
 
 vO vO i-H CM 00 00 ^ vO t*> 
 
 cm oo to .— i rt oo oo cm to cn oo so \D 
 
 LO O 
 
 to cn i— i cm oo oo cn to rj- cn lo i— ivo 
 
 V0 o 
 
 «tf \D 
 
 t0 "* 
 
 CN 
 
 lo cn cn cn cm looo(n cn r^ 
 
 LO O 
 
 t^-oor^crii— ii— i^ovOtNi— ilo r^ 
 
 LO O 
 
 i— I Ch 00 rt to i— l f*> to i— 1 O \D 
 
 "tf o 
 
 to 
 
 o 
 
 o 
 
 to 
 
 o 
 
 to 
 
 o 
 
 o 
 
 "3- 
 
 o 
 
 o 
 
 CM 
 
 o 
 
 ■— icMto-3-LOvor-^-ooenOi— i 
 
 cn to Tt aj g 
 
 i— I I— I .—I +-) f-| +J 
 
 O O W 
 
 H Z UJ 
 
 LO 
 
 r- 
 
 
 ^o 
 
 y 
 
 CD 
 
 CD 
 
 a> 
 
 4-> 
 
 o 
 
 LW 
 
 rt 
 
 c 
 
 ■H 
 
 u 
 
 o 
 
 o 
 
 Ph 
 
 M 
 
 o 
 
 +-> 
 
 
 <u 
 
 en 
 
 rt 
 +J 
 
 03 
 Q 
 
 vO 
 
 0) 
 
 (■H 
 3 
 
 M 
 
 •H 
 
20 
 
 became available) and were completed on April 21. The April estimates were 
 completed on April 21. The April estimates were completed at a more leis- 
 urely pace. Figures 17 and 18 are examples of weekly precipitation esti- 
 mates. Figure 19 shows the number of consecutive days in April for which no 
 precipitation was estimated at various USSR grid points. This statistic is 
 important for agricultural planning. 
 
 Precipitation data over the USSR were available for verification, although 
 much was of questionable quality. No Chinese precipitation data were 
 available. 
 
 For the 4-week period, March 2-29, 1975, total observed precipitation was 
 determined for each reporting point in the parts of the USSR under study. 
 The total for each reporting point was compared with the 4-week estimate at 
 the nearest grid point. The mean absolute error was 55% of the mean observed 
 precipitation. Bias, however, was virtually zero. 
 
 The technique was tested over the central United States for April 1976. 
 Daily estimates were made for 547 rain reporting points in the 24 states 
 lying between the Rockies and the Appalachians. Estimates were made for 
 individual stations rather than for grid points to make use of excellent 
 climatological normals and daily reports afforded by the stations used in 
 the sample. 
 
 Because of picture-taking times of the polar-orbiting satellite, NOAA-4, 
 the estimates cover the period from 8 p.m. to 8 p.m., local time, approxi- 
 mately. If all reporting points took observations at 8 p.m., local time, 
 verification of the technique would be greatly facilitated. Unfortunately 
 for this study, only 1 to 2% of the readings received by the Environmental 
 Data Service for publication in the Clvraatologieal Data are taken at this 
 time. Therefore, in selecting the verification sample, the primary data 
 sources chosen are the National Weather Service and Federal Aviation Admini- 
 stration stations, which take rainfall measurements at midnight local stand- 
 ard time. The secondary source is a group of climatological and river- 
 rainfall stations taking observations between 4 p.m. and 8 p.m., local 
 standard time; these stations fill the gaps in the primary network. In 
 areas devoid of such points, a few stations that take morning or early 
 afternoon observations have been added to the sample. However, the time 
 discrepancies thus introduced perhaps may have been too great a price to pay 
 for the increase in geographical coverage. 
 
 Of the several verification checks applied to the United States estimates 
 for April 1976, all but one show more accurate estimates for the primary 
 source stations (NWS and FAA) than for the sample as a whole. 
 
 The grand total of daily errors (disregarding sign) for all stations for 
 the entire month has been divided by the grand total of observed precipita- 
 tion. The ratio for the entire sample is 1.00, and for the primary source 
 stations, 0.96. Similar ratios have been calculated for each station. The 
 median ratio for the entire sample is 0.98, and for the primary source 
 stations, 0.95. 
 
2] 
 
 a 
 
 ^ 
 
 p, 
 
 
 
 CC 
 O 
 
 S 03 
 
 
 
 
 
 -" 
 
 o 
 
 m 
 
 
 In 
 
 [ri 
 
 
 
 
 
 rj 
 
 ^T 
 
 UD 
 
 CO 
 
 8 
 
 r-g 
 
 
 LU 
 O 
 
 a 
 
 
 o 
 
 ~ 
 
 ro 
 
 f* 
 
 rj 
 
 fj 
 
 O 
 
 S 
 
 ° 
 
 o 
 
 CD 
 
 o 
 
 8 
 
 - 
 
 o 
 
 r- 
 
 in 
 
 m 
 
 
 
 'X) 
 
 o 
 
 tfi 
 
 o 
 
 N 
 
 p* 
 
 t- 
 
 KJ 
 
 M 
 
 
 3 
 
 a 
 
 q 
 
 to 
 
 o 
 
 ~s~ 
 
 - 
 
 h- 
 
 ■CN|- 
 
 - cq 
 
 U3 
 M 
 
 CM 
 
 q 
 
 
 O 
 
 '+H 
 
 4h 
 
 
 
 
 • 
 
 
 o 
 
 c/> 
 
 > 
 
 ~c 
 
 •H 
 
 4J 
 
 % 
 
 -3 
 
 3 
 
 O 
 
 ^H 
 
 fn 
 
 'J 
 
 TJ 
 
 c 
 
 c 
 
 • H 
 
 3 
 
 
 X 
 
 * 
 
 
 LO 
 
 c 
 
 1 
 
 ■H 
 
 CT> 
 
 
 r— 1 
 
 t/1 
 
 
 O 
 
 - 
 
 4-J 
 
 cr, 
 
 d 
 
 i-H 
 
 6 
 
 
 ■H 
 
 o 
 
 fJ 
 
 +-S 
 
 V) 
 
 
 o 
 
 to 
 
 
 ^H 
 
 c 
 
 
 o 
 
 1— 1 
 
 •H 
 
 •H 
 
 «-) 
 
 F-l 
 
 cti 
 
 ft 
 
 4-> 
 
 < 
 
 •H 
 
 
 Ph 
 
 FH 
 
 ■H 
 
 o 
 
 U 
 
 U-i 
 
 O 
 
 
 'm 
 
 
 Ph 
 
 
 >. 
 
 
 1 — 1 
 
 
 -^ 
 
 
 o 
 
 
 o 
 
 
 o 
 - 
 
22 
 
 130E 
 
 50N 
 
 120E 
 
 25 N 
 
 100E 
 
 110E 
 
 Figure 18. --Weekly precipitation estimates in hundredths of an inch 
 for China for April 13 to 19', 1975, inclusive. 
 
23 
 
 4-> 
 
 ■H 
 
 Ph 
 
 ■H 
 O 
 
 o 
 
 Ph 
 
 o 
 
 X 
 
 o 
 
 fH 
 
 
 o 
 
 in 
 
 4n 
 
 +J 
 
 
 (3 
 
 LO 
 
 ■H 
 
 r^ 
 
 O 
 
 G> PL, 
 I— I 
 
 13 
 
 ■ H 
 
 'M 
 
 h 
 
 M 
 
 ft 
 
 
 < 
 
 OJ 
 
 
 CO 
 
 C 
 
 to 
 
 •H 
 
 3 
 
 (/) 
 
 •M 
 
 >. 
 
 cd 
 
 rt 
 
 
 13 
 
 "3 
 
 
 O 
 
 o 
 
 fJ 
 
 > 
 
 m 
 
 •H 
 
 s 
 
 +J 
 
 ■H 
 
 3 
 
 4-J 
 
 O 
 
 t/} 
 
 O 
 
 O 
 
 t/1 
 
 
 c 
 
 V) 
 
 o 
 
 ti 
 
 •J 
 
 5 
 
 4h 
 
 
 o 
 
 
 M 
 
 
 o 
 
 
 rC 
 
 
 o 
 
 SO 
 
 •H 
 
24 
 
 The above ratios are based on daily errors, a rigorous test. The ratio 
 of monthly error to total monthly precipitation is far less exact, but 
 useful. The grand total of monthly error divided by the grand total of 
 monthly precipitation gives, for the full sample, a ratio of 0.46 (compared 
 with 0.55 for the USSR in March 1975). The ratio for the primary source 
 stations is 0.49; this is the sole statistic favoring the entire sample. 
 
 A categorical rain-no- rain test has been applied to the estimates. The 
 results are shown in table 1 for the full sample, and in table 2 for the 
 primary source stations. The primary source estimates outscore the entire 
 sample estimates 75% to 73% in percent correct, 0.47 to 0.42 in skill score, 
 0.49 to 0.4 5 in threat score, 0.54 to 0.50 in post agreement, 0.83 to 0.80 
 in prefigurance, and 1.53 to 1.58 in bias. For the entire sample, rain was 
 estimated for 80% of the actual rain cases, and no rain was estimated for 
 70% of the non-rain cases. The comparable figures for the primary source 
 stations are 83% and 71%. 
 
 Inspection of the day-by-day estimated and observed precipitation for each 
 of the 547 stations in the sample provides a better understanding of the 
 strengths and weaknesses of the technique than can be had from the above 
 statistics. Table 3 shows two of the most successful cases, table 4 two of 
 the least successful, and table 5 two in the middle of the accuracy spectrum. 
 
 The most serious failures of the technique were due to the high frequency 
 of thunderstorms over most of the area under study during April 1976. The 
 technique is tailored to less convective precipitation regimes, and should 
 perhaps be restricted to the cooler months from October to March. (Even 
 March 1976 yielded a record-breaking count of 211 tornadoes with copious 
 convective rains over the United States. The normal count is in the low 
 forties, however.) Goliad, Texas, highlights the problem (table 4). The 
 observed amounts are much greater than the technique can estimate. In 
 addition, thunderstorms can originate, mature, and decay well within the 
 time interval between successive 12-hour pictures. No doubt this happened 
 at Goliad on April 7 and 19. In contrast, the overestimates for parts of 
 Louisiana and Arkansas may be due to mistaking thick debris from Texas and 
 Oklahoma thunderstorms for rain-producing clouds. 
 
 FUTURE INVESTIGATIONS 
 
 It should be possible to increase the accuracy of the method by developing 
 or adapting various models of precipitation patterns. Oliver (personal 
 communication 1974) has suggested a model that should apply to most middle 
 latitude storms. The essentials of the model (Fig. 20) are generally not 
 difficult to identify in either visible or infrared satellite cloud imagery. 
 The location of the jet axis has been described in a number of papers. Its 
 role in the model is crucial. Normally the heaviest precipitation occurs in 
 the area where the jet crosses the frontal band near the point of occlusion. 
 Significant precipitation is also found in the cold frontal band not far 
 south of the point of occlusion. Little or none occurs under that part of 
 the jet which has a northerly component (that is, on the east side of the 
 upper level ridge line) . 
 
Table 1 . --Categorical rain-no rain verification of esti- 
 mates for 547 stations over central United States. 
 
 25 
 
 Total cases (T) = 16410 
 
 Total correct (7?) = 11961 
 
 Percent correct = 73% 
 
 Expected correct 
 by chance (E) 
 
 = (q+fc)(a+e) + (c+d)(fe+d) = (4519) (7154)+(11891) (9256) 
 T 16410 
 
 „,.,, R-E 11961-8677 .. 
 Skill score = ^ = 16410 _ 8677 = -^ 
 
 3612 
 
 
 
 
 Estimated 
 
 
 
 
 Rain 
 
 None 
 
 Total 
 
 
 Rain 
 
 3612(a) 
 
 907(&) 
 
 4519(a) 
 
 
 None 
 
 3542(e) 
 
 8349(d) 
 
 11891(0+0") 
 
 .0 
 
 o 
 
 Total 
 
 7154(a+e) 
 
 9256 (W) 
 
 16410 (a+&+e+d) 
 
 8677 
 
 Threat score = 
 
 Post agreement 
 
 3612+907+3542 
 3612 
 
 = .45 
 
 7154 
 
 ,50 
 
 Prefigurance = ,,., „ = .80 
 
 7154 
 4519 
 
 4519 
 
 = 1.58 
 
 Table 2. --Categorical rain-no rain verification of esti- 
 mates for first order and airport stations (162) . 
 
 
 
 
 
 Estimated 
 
 
 
 
 
 Rain 
 
 None 
 
 Total 
 
 Total cases (T) = 4860 
 
 -Q 
 
 Rain 
 
 1185(a) 
 
 243(fo) 
 
 1428 (a+Z?) 
 
 Total correct (/?) = 3621 
 
 > 
 
 None 
 
 996(e) 
 
 2436(d) 
 
 3432 (e+d) 
 
 Percent correct = 75°6 
 
 X> 
 
 
 
 
 
 Expected correct 
 
 o 
 
 Total 
 
 2181 (a+e) 
 
 2679 (£>+ d) 
 
 4 860(a+&+e+d) 
 
 by chance (E) 
 
 
 
 
 
 
 (1428) (2181)+(3432) (2679) 
 4860 
 
 Skill score = 
 Threat score = 
 Post agreement 
 Prefigurance = 
 
 R-E _ 3621-2533 
 T-E ~ 4860-2533 
 
 1185 = 
 1185+243+996 
 
 1185 
 
 2533 
 
 .47 
 
 .49 
 
 2181 
 
 I is.. 
 1428 
 
 .54 
 
 3ias = 
 
 2181 
 1428 
 
 = 1.53 
 
26 
 
 Table 3. --Estimated and observed daily rainfall for selected 
 stations in the United States, April 1976. (Best cases.) 
 
 1976 
 
 Youngstown, Ohio 
 Est Obs E-0 
 
 Rockford, Illinois 
 Est Obs E-0 
 
 ril 1 
 
 0.07 
 
 0.10 
 
 -0.03 
 
 2 
 
 .04 
 
 .04 
 
 
 
 3 
 
 
 
 
 
 
 
 4 
 
 .07 
 
 .17 
 
 - .10 
 
 5 
 
 
 
 
 
 
 
 6 
 
 .04 
 
 .01 
 
 + .03 
 
 7 
 
 
 
 .04 
 
 -.04 
 
 8 
 
 
 
 T 
 
 
 
 9 
 
 
 
 
 
 
 
 10 
 
 
 
 
 
 
 
 11 
 
 .04 
 
 .03 
 
 + .01 
 
 12 
 
 
 
 
 
 
 
 13 
 
 
 
 
 
 
 
 14 
 
 
 
 
 
 
 
 15 
 
 
 
 
 
 
 
 16 
 
 .04 
 
 .04 
 
 
 
 17 
 
 .04 
 
 
 
 + .04 
 
 18 
 
 
 
 
 
 
 
 19 
 
 
 
 
 
 
 
 20 
 
 
 
 
 
 
 
 21 
 
 .29 
 
 .38 
 
 - .09 
 
 22 
 
 .11 
 
 .32 
 
 -.21 
 
 23 
 
 .04 
 
 
 
 + .04 
 
 24 
 
 .18 
 
 .17 
 
 + .01 
 
 25 
 
 .40 
 
 .26 
 
 + .14 
 
 26 
 
 .04 
 
 .08 
 
 -.04 
 
 27 
 
 
 
 T 
 
 
 
 28 
 
 
 
 
 
 
 
 29 
 
 
 
 
 
 
 
 30 
 
 
 
 
 
 
 
 Tot?} 
 
 1.40 
 
 1.64 
 
 0.78* 
 
 Ra t i o : 
 
 0.48 
 
 
 -0.24# 
 
 0.04 
 
 T 
 
 + 0.04 
 
 
 
 
 
 
 
 
 
 .04 
 
 - .04 
 
 
 
 .03 
 
 - .03 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .04 
 
 .16 
 
 - .12 
 
 .08 
 
 .05 
 
 + .03 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .35 
 
 .86 
 
 - .51 
 
 .12 
 
 T 
 
 + .12 
 
 .08 
 
 .04 
 
 + .04 
 
 .12 
 
 .02 
 
 + .10 
 
 
 
 
 
 
 
 .19 
 
 .51 
 
 -.32 
 
 .35 
 
 .35 
 
 
 
 
 
 
 
 
 
 .38 
 
 .50 
 
 - .12 
 
 .62 
 
 .75 
 
 - .13 
 
 .35 
 
 .17 
 
 + .18 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .08 
 
 .12 
 
 - .04 
 
 2.80 
 
 3.60 
 
 1.82* 
 
 atio: 
 
 0.51 
 
 -0.80# 
 
 *Total of daily errors, disregarding sign. 
 
 #Algebraic sum of daily errors, which equals total estimated 
 minus total observed. 
 
Table 4. --Estimated and observed daily rainfall for selected 
 stations in the United States, April 1976. (Worst cases.) 
 
 11 
 
 1976 
 
 Jennings, Louisiana 
 Est Obs E-0 
 
 Goliad, Texas 
 Est Obs E-0 
 
 April 1 
 
 
 
 
 
 
 
 2 
 
 1! 
 
 
 
 
 
 3 
 
 (1 
 
 
 
 
 
 4 
 
 
 
 (1 
 
 ii 
 
 5 
 
 .87 
 
 
 
 + .87 
 
 <: 
 
 .27 
 
 .07 
 
 + .20 
 
 7 
 
 .38 
 
 
 
 + .38 
 
 8 
 
 
 
 .05 
 
 - .05 
 
 9 
 
 
 
 
 
 
 
 10 
 
 
 
 
 
 
 
 1 i 
 
 
 
 
 
 
 
 12 
 
 '.i 
 
 
 
 
 
 13 
 
 .33 
 
 
 
 + .33 
 
 14 
 
 
 
 .02 
 
 - .02 
 
 15 
 
 .16 
 
 
 
 + .16 
 
 16 
 
 .11 
 
 
 
 + .11 
 
 , 7 
 
 
 
 
 
 
 
 IS 
 
 (! 
 
 
 
 
 
 l ! '< 
 
 
 
 
 
 
 
 20 
 
 .27 
 
 
 
 + .27 
 
 21 
 
 
 
 .42 
 
 - .42 
 
 22 
 
 
 
 
 
 
 
 25 
 
 
 
 
 
 (i 
 
 24 
 
 .27 
 
 
 
 + .27 
 
 25 
 
 .11 
 
 .30 
 
 - .19 
 
 26 
 
 (' 
 
 
 
 
 
 27 
 
 
 
 
 
 
 
 28 
 
 
 
 li 
 
 
 
 29 
 
 .49 
 
 
 
 + .49 
 
 50 
 
 .16 
 
 
 
 + .16 
 
 Total 
 
 3.42 
 
 0.86 
 
 3.92* 
 
 Ratio : 
 
 4.56 
 
 
 +2.56# 
 
 0.03 
 
 
 
 
 + 0.03 
 
 .13 
 
 
 
 
 + .13 
 
 ii 
 
 
 1) 
 
 
 
 .42 
 
 
 .31 
 
 + .11 
 
 .03 
 
 1 
 
 .18 
 
 -1.15 
 
 .16 
 
 
 
 
 + .16 
 
 .18 
 
 
 .08 
 
 + .10 
 
 
 
 1 
 
 .75 
 
 -1.75 
 
 .03 
 
 
 
 
 + .03 
 
 
 
 
 
 
 
 
 
 
 
 n 
 
 o 
 
 .05 
 
 
 T 
 
 + .05 
 
 .10 
 
 
 .17 
 
 - .07 
 
 .03 
 
 
 
 
 + .03 
 
 .08 
 
 
 
 
 + .08 
 
 (i 
 
 
 .39 
 
 - .39 
 
 () 
 
 
 
 
 o 
 
 .23 
 
 5 
 
 .00 
 
 -4.77 
 
 
 
 3 
 
 .35 
 
 -3.35 
 
 .16 
 
 
 .72 
 
 - .56 
 
 
 
 
 
 
 
 
 .03 
 
 
 
 
 + .03 
 
 .16 
 
 
 
 
 + .16 
 
 .42 
 
 
 
 
 + .42 
 
 .13 
 
 
 .15 
 
 -.02 
 
 
 
 
 
 
 
 
 I) 
 
 
 
 
 
 
 .10 
 
 
 .01 
 
 + .09 
 
 .31 
 
 1 
 
 .11 
 
 - .80 
 
 .03 
 
 
 .01 
 
 + .02 
 
 2.81 
 
 14 
 
 .23 
 
 14.30* 
 
 .atio : 
 
 1 
 
 .00 
 
 -11.42# 
 
 *Total of daily errors, disregarding sign. 
 
 #Algebraic sum of daily errors, which equals total estimated 
 minus total observed. 
 
28 
 
 Table 5. --Estimated and observed daily rainfall for selected sta- 
 tions in the United States, April 1976. (Average cases.) 
 
 1976 
 
 Ridgeland, Wisconsin 
 Est Obs E-0 
 
 Cherokee, Oklahoma 
 Est Obs E-0 
 
 April 1 
 
 
 
 
 
 
 
 2 
 
 
 
 
 
 
 
 3 
 
 .09 
 
 T 
 
 + .09 
 
 4 
 
 
 
 
 
 
 
 5 
 
 
 
 
 
 
 
 6 
 
 
 
 
 
 
 
 7 
 
 
 
 
 
 
 
 8 
 
 
 
 
 
 
 
 9 
 
 .03 
 
 
 
 + .03 
 
 10 
 
 .11 
 
 .22 
 
 - .11 
 
 1 1 
 
 
 
 
 
 
 
 12 
 
 
 
 
 
 
 
 13 
 
 
 
 
 
 
 
 14 
 
 .03 
 
 .69 
 
 - .66 
 
 15 
 
 .03 
 
 
 
 + .03 
 
 L6 
 
 .09 
 
 
 
 + .09 
 
 17 
 
 .20 
 
 .04 
 
 + .16 
 
 18 
 
 
 
 .90 
 
 - .90 
 
 19 
 
 
 
 
 
 
 
 20 
 
 .09 
 
 
 
 + .09 
 
 21 
 
 .26 
 
 .26 
 
 
 
 11 
 
 .03 
 
 .01 
 
 + .02 
 
 2 3 
 
 .09 
 
 .24 
 
 - .15 
 
 24 
 
 .40 
 
 .27 
 
 + .13 
 
 2 5 
 
 .11 
 
 
 
 + .11 
 
 26 
 
 
 
 
 
 
 
 27 
 
 
 
 
 
 
 
 28 
 
 
 
 
 
 
 
 29 
 
 
 
 
 
 
 
 30 
 
 
 
 T 
 
 
 
 Total 
 
 1.56 
 
 2.63 
 
 2.57* 
 
 Ra t i o : 
 
 0.98 
 
 
 -1.07# 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .21 
 
 T 
 
 + .21 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .30 
 
 .02 
 
 + .28 
 
 .03 
 
 
 
 + .03 
 
 
 
 .46 
 
 - .46 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .09 
 
 T 
 
 + .09 
 
 .18 
 
 .69 
 
 - .51 
 
 .06 
 
 .60 
 
 - .54 
 
 
 
 
 
 
 
 .15 
 
 
 
 + .15 
 
 .30 
 
 .59 
 
 -.29 
 
 
 
 
 
 
 
 
 
 
 
 
 
 .06 
 
 
 
 + .06 
 
 .09 
 
 .20 
 
 - .11 
 
 
 
 
 
 
 
 .06 
 
 T 
 
 + .06 
 
 .09 
 
 T 
 
 + .09 
 
 .36 
 
 .91 
 
 - .55 
 
 
 
 
 
 
 
 .12 
 
 .34 
 
 - .22 
 
 2.10 
 
 3.81 
 
 3.65* 
 
 atio : 
 
 0.96 
 
 -1.71* 
 
 *Total of daily errors, disregarding sign. 
 
 #Algebraic sum of daily errors, which equals total estimated 
 minus total observed. 
 
29 
 
 VERY LITTLE PRECIPITATION 
 EAST OF RIDGE LINE 
 
 P = POINT OF OCCLUSION. TIGHT TEMPERATURE 
 GRADIENT AT "P" UNDER JET. ALTOSTRATUS 
 MOVING NNW ACCROSS JET AXIS ENCOUNTERS 
 MAXIMUM UPWARD MOTION; THEREFORE 
 HEAVIEST PRECIPITATION. 
 
 Figure 20. --Middle-latitude storm model showing zones of greatest 
 intensity of precipitation (from Oliver) . 
 
 Smith and Younkin (1972) have shown that for "digging" polar jets over the 
 central United States, heavy rainfall tends to occur in an ellipsoidal pattern 
 east of the jet stream. Their composite model for forecasting 12-hour heavy 
 rainfall should be tested, both over the central United States and other parts 
 of the globe, for possible use in this technique. Other investigations by 
 Younkin and his collaborators (1968, 1970) and by Fawcett and Saylor (1965) 
 should be tested in the same way. 
 
 ACKNOWLEDGMENTS 
 
 The author wishes to thank Vincent J. Oliver and Dr. Norton D. Strommen 
 for valuable suggestions for improving the paper; Paul Lehr for his editorial 
 review; and Peg Follansbee, his personal secretary, who helped so much. 
 
30 
 
 REFERENCES 
 
 Browne, Richard F., and Younkin, Russell J., "Some Relationships 
 Between 850-Millibar Lows and Heavy Snow Occurrences Over the 
 Central and Eastern United States," Monthly Weather Review, 
 Vol. 98, No. 5, May 1970, pp. 399-401. 
 
 Fawcett, E. B. , and Saylor, H. K. , "A Study of the Distribution of 
 Weather Accompanying Colorado Cyclogenesis," Monthly Weather 
 Review, Vol. 93, No. 6, June 1965, pp. 359-367. 
 
 Goree, Paul A., and Younkin, Russell J., "Synoptic Climatology of 
 Heavy Snowfall Over the Central and Eastern United States," 
 Monthly Weather Review, Vol. 94, No. 11, November 1966, pp. 663-668 
 
 Smith, Warren, and Younkin, Russell J., "An Operationally Useful 
 
 Relationship Between the Polar Jet Stream and Heavy Precipitation," 
 Monthly Weather Review, Vol. 100, No, 6, June 1972, pp. 434-440. 
 
 Oliver, Vincent J., National Environmental Satellite Service, 
 Washington, D.C., 1974, personal communication. 
 
 Younkin, Russell J., "Circulation Patterns Associated With Heavy 
 Snowfall Over the Western United States," Monthly Weather Review, 
 Vol. 96, No. 12, December 1968, pp. 851-853. 
 
(Continued from inside front cover) 
 
 NESS 59 Use of Geostationary-Satellite Cloud Vectors to Estimate Tropical Cyclone Intensity. Carl. 0. 
 Erickson, September 1974, 37 pp. (C0M-74-11762/AS) 
 
 NESS 60 The Operation of the NOAA Polar Satellite System. Joseph J. Fortuna and Larry N. Hambrick, 
 November 1974, 127 pp. (COM-75-10390/AS) 
 
 NESS 61 Potential Value of Earth Satellite Measurements to Oceanographic Research in the Southern 
 Ocean. E. Paul McClain, January 1975, 18 pp. (COM-75-10479/AS) 
 
 NESS 62 A Comparison of Infrared Imagery and Video Pictures in the Estimation of Daily Rainfall From 
 Satellite Data. Walton A. Follansbee and Vincent J. Oliver, January 1975, 14 pp. (COM-75- 
 10435/AS) 
 
 NESS 63 Snow Depth and Snow Extent Using VHRR Data From the N0AA-2 Satellite. David F. McGinnis, Jr., 
 John A. Pritchard, and Donald R. Wiesnet, February 1975, 10 pp. (COM-75-10482/AS) 
 
 NESS 64 Central Processing and Analysis of Geostationary Satellite Data. Charles F. Bristor (Editor), 
 March 1975, 155 pp. (COM-75-10853/AS) 
 
 NESS 65 Geographical Relations Between a Satellite and a Point Viewed Perpendicular to the Satellite 
 Velocity Vector (Side Scan). Irwin Ruff and Arnold Gruber, March 1975, 14 pp. (COM-75-10678/AS) 
 10678/AS) 
 
 NESS 66 A Summary of the Radiometric Technology Model of the Ocean Surface in the Microwave Region. 
 John C. Alishouse, March 1975, 24 pp. (COM-75-10849/AS) 
 
 MESS 67 Data Collection System Geostationary Operational Environmental Satellite: Preliminary Report. 
 Merle L. Nelson, March 1975, 48 pp. (COM-75-10679/AS) 
 
 NESS 68 Atlantic Tropical Cyclone Classifications for 1974. Donald C. Gaby, Donald R. Cochran, James 
 B. Lushine, Samuel C. Pearce, Arthur C. Pike, and Kenneth 0. Poteat, April 1975, 6 pp. (COM-75- 
 1-676/ AS) 
 
 NESS 69 Publications and Final Reports on Contracts and Grants, NESS-1974. April 1975, 7 pp. (COM- 
 75-10850/AS) 
 
 NESS 70 Dependence of VTPR Transmittance Profiles' and Observed Radiances on Spectral Line Shape Parame- 
 ters. Charles Braun, July 1975, 17 pp. 
 
 NESS 71 Nimbus-5 Sounder Data Processing System, Part II: Results. W. L. Smith, H. M. Woolf, C. M. 
 Hayden, and W. C. Shen. July 1975, 102 pp. 
 
 NESS 72 Radiation Budget Data From the Meteorological Satellites, ITOS 1 and NOAA 1. Donald H. 
 Flanders and William L. Smith, August 1975, 22 pp. 
 
 NESS 73 Operational Processing of Solar Proton Monitor Data. Stanley R. Brown, September 1975. (Re- 
 vision of NOAA TM NESS 49), 15 pp. 
 
 NESS 74 Monthly Winter Snowline Variation in the Northern Hemisphere from Satellite Records, 1966x75. 
 Donald R. Wiesnet and Michael Matson, November 1975, 21 pp. (PB248437) 
 
 NESS 75 Atlantic Tropical and Subtropical Cyclone Classifications for 1975. D. C. Gaby, J. B. Lushine, 
 B. M. Mayfield, S. C. Pearce, and K. 0. Poteat, March, 1976, 14 pp. 
 
 NESS 76 The Use of the Radiosonde in Deriving Temperature Soundings From the Nimbus and NOAA Satellite 
 Data. Christopher M. Hayden, April 1976, 21 pp. (PB-256755) 
 
 NESS 77 Algorithm for Correcting the VHRR Imagery for Geometric Distortions Due to the Earth's Curva- 
 ture and Rotation. Richard Legeckis and John Pritchard, April 1976, 30 pp. 
 
 NESS 78 Satellite Derived Sea-Surface Temperatures From NOAA Spacecraft. Robert L. Brower, Hilda S. 
 Gohrband, William G. Pichel, T. L. Signore, and Charles C. Walton, in press, 1975. 
 
 NESS 79 Publications and Final Reports on Contracts and Grants, 1975. NESS, June 1976. 
 
 NESS 80 Satellite Images of Lake Erie Ice: January-March 1975. Michael C. McMillan and David Forsyth, 
 June 1976. 
 
ADDDD173bl37fl 
 
 ot- UT| o^ 
 
 NOAA--S/T 76-2464 
 
 ">6-«1 fe ®