Comparison between physical and Statistical Methods for Estimation of PMP in Southwest basins of Iran

A.M, Noorian1 and E. Fattahi2 Email: amnoorian@yahoo.com ebfat2002@yahoo.com 1, 2: Atmospheric Science and Meteorological Research Center (ASMERC) Iran.

Abstract
The probable maximum precipitation (PMP) is the greatest depth of precipitation for a given duration that is physically possible over a given size storm area at a particular geographical location at a certain time of the year. Hydrologists use a PMP magnitude together with its spatial and temporal distributions for the catchments of a dam to calculate the probable maximum flood (PMF). In this study the physical or maximization of storm method has been compared with statistical method (e. g. the Hershfilds) for calculate PMP in southwest regions of Iran. In order to estimate of probable maximum precipitation first collection meteorological data including precipitation; dew point temperature; wind speed were collected. In this connection synoptic maps at appropriate time scales at meteorological stations in southwest catchments and neighboring basins were study. Then major storms (extreme) including last year storms of various durations were selected and Depth-Area-Duration (DAD) curves for all selected storms extracted. By using physical Method, PMP estimations were obtained at different locations. These estimations PMP values can be applied to 1000, 5000 and 10,000 km2 areas. In this study also PMP estimations were obtained by statistical analysis (Hershfilds Method) of the series of annual maximum 24h precipitation amounts. We found that PMP estimates by statistical method are well comparable with values of obtained by the physical method for different durations. Results also shows that limited transposition of statistical methods gives higher estimates, in comparisons with physical method.

Keyword: Probable Maximum Precipitation, (pmp), Hirschfield's method, Depth-Area-Duration (DAD), Dew point Temperature.

1. Introduction
According to the world meteorological organization definition (WMO,1986) the probable maximum precipitation is defined as theoretically the greatest depth of precipitation for a given duration that is physically possible over a given size storm area at a particular geographical location at a particular time of year. Hydrologists use a PMP magnitude to gather with its spatial and temporal distributions, for the catchments of a dam to calculate the Probable Maximum Flood (PMF).The PMF is one of a range of conceptual flood events used in the design of hydrological structures. Since the 1950 a number of methods have been developed for estimating PMP in USA and Australia. The Hydro meteorological Report (HMR) published by the us department of commerce is HMR 59 (Corrigan et .al., 1999) for California in 1999. The generalized tropical method, for regions of Australia affected by tropical storms; was developed during the 1970 (Kennedy, 1982). The generalized short duration method, for small-area PMP estimation, was developed in the early 1980, (Bureau of meteorology, 1985. Collier (1996) estimated PMP value for the UK using convective storm mode. A localized study on the statistical estimation of PMP (Hershfilds method) was carried out by Desa et. al. (2001) for the state of Selangor. Estimates of PMP in Iran were based on maximization method. The purpose of this paper is to provide a method that can be used to estimates of PMP for catchments area up to 1000Km2.

2. Methodology
The most common methods to estimates PMP are storm maximization (Hydro meteorological) approach and statistical approach Hershfield method. Storm maximization and transpositation method requires more sitespecific data and thus provides more reliable estimate than other methods. Where sitespecific data are not available statistical method (Hershfield method) can be applied that requires data for annual maximum

rainfall series in the region for required storm durations. Factors that in fluency calculations of PMP values are rainfall, dew point, temperature, wind speed, temperature and pressure.

Figure (1) Southwest basins of Iran

2.1. Physically or maximization method 2.1.1. Depth-Area-Duration

An important step probable maximum estimates is the analysis of major storms. This analysis includes; collection precipitation data from various sources. The objective of the storm analysis was to obtain Deph-Area-Duration (DAD) information upon which to base PMP estimates as well as generalized relations for other areas or other basin with similar climate and topographic characteristics. The first step in order to development of DAD relations requires that rainfall amounts to all areas in the storms. Then point precipitation amounts interpolated by assigning a particular precipitation gauge to a region. Estimation of spatial distribution of rainfall is one of the basic steps in PMP studies. In this study in order to interpolation were used kriging methods.

2.1.2. Maximum of storms humidity and dew point factor

Storm maximization, it is assumed that rainfall can be determined from the product of an available moisture and storm mechanism. After selection of storms and estimation of mean rainfall depth for each sub-basin, it was necessary to estimate maximum humidity source in order to maximize selected storms. Storm moisture maximization factor was determined using the surface dew point temperature, in conjunction with an assumed saturated atmosphere above surface level. Surface dew point was used as a measure of moisture potential for server storms because it is the critical factor for server storm development in small areas. Maximum dew point for any location is chosen as the highest value persisting for 12-hour duration. In this study recorded surface dew points 7 station could be considered as representative of in flow humidity source. There fore, to calculate the maximum in flow humidity to storm, it is needed to investigate maximum persisting 12-hour dew point duration for statistical period of 7 station and also dew point values of those station at storm event. For investigation of maximum persisting 12-hour dew point duration in statistical period, the 10-day recorded data for each year were

extracted and then maximum persisting 12-hour dew point values for each 10-day period were calculated with 50-year return period by using frequency analysis. Having used the mean monthly pressure data recorded at each station were transferred to 1000 millibar pressure level such that they can be compared at different elevations. The calculated dew point temperature at the 7 station during all storm events and maximum 12-hour persisting condition were reduced to equivalent mean sea level (Msl, i.e.1000 millibar air pressure) dew point temperatures, using figure 2.

Figure(2): Peseoud-adiabatic chart for dew point reducation to 1000 hpa at high zero
In next step, variation of 10-day maximum 12-hour duration dew point of each station with 50-year return period (at level 1000 millibar) are used the corresponding typical curves were plotted and then by reading points from those curves, the humidity coefficients were estimated. Mentioned curves are shown in figure (2). Usually recorded dew points in stormy days of a 12-hour period, which have the biggest values, are selected to investigate 12-hour duration dew point at the selected storms. Then the smallest of these values during storm event were used as maximum persisting 12-hour duration of dew point. These values were also transferred to 1000 millibar level by using stations air pressures data during storm event to become comparable with each other. Then, the humidity coefficient calculated by using equation1 FM=Wm/Ws (1) in this equation Wm is maximum precipitable water at 1000 to 200 millibar levels which can be obtained on the basis of maximum 12-hour duration dew point with 50-year return period in a simultaneous period with storm. Ws= maximum perceptible water at 1000 to 200 millibar levels which can be obtained on the basis of maximum 12-hour duration dew point in a simultaneous period with storm. The relevant values of dew point temperature and humidity coefficients are presented in table (1) and (2).

Table (2)

the Moisture and Wind Maximization Factors for Some Selected Storms

Storms Date

Synoptic Station

Maximum persisting 12hr dew points in 1000 mb level (in the storm time) ()

Precipitable water(mm)

Maximum persisting 12hr dew point in 1000 mb level (100 year return period) ()

Precipitable water(mm)

Moisture factor

Maximum persisting 12hr wind (knot)

Maximum persisting 12hr wind(50 year return period) (knot)

Wind Factor

PMP Factor

Abadan

16.9

39.6

19.4

49.6

1.25

18.0

21.6

1.2

Ahvaz

15.7

35.1

20.9

55.6

1.58

18

26.1

1.45

Mahshar

14.8

32.4

19.1

48.4

1.26

22

31.4

1.42

30 Mar 1998

Omidyeh

16.6

38.4

19.1

48.4

1.26

18

20.5

1.14

Dezful

17

40

20.4

53.6

1.34

18

20.8

1.15

1.34 Averag e

1.27

1.70

Table (1) the Moisture

and Wind Maximization Factors for Some Selected Storms

Storms Date

Synoptic Station

Maximum persisting 12hr dew points in 1000 mb level (in the storm time) ()

Precipitable water(mm)

Maximum persisting 12hr dew point in 1000 mb level (100 year return period) ()

Precipitable water(mm)

Moisture factor

Maximum persisting 12hr wind (knot)

Maximum persisting 12hr wind(50 year return period) (knot)

Wind Factor

PMP Factor

Abadan

20.1

52.0

22.0

62.0

1.2

18.0

21.0

1.16

Ahvaz

17.4

41.6

21.1

57.5

1.38

21

23.5

1.2

Bushehr 23-26 Nov 1994

17.1

40.4

21.2

58.0

1.43

15

26.1

1.74

Mahshar

19

48

20.9

56.5

1.18

18

31.4

1.74

Omidyeh

17.2

40.8

23.4

64.4

1.57

17

21.8

1.28

Bushehr Coastal

19.6

50.4

21.8

61

1.21

20

32.5

1.62

Aghajari

17.7

42.8

22.4

64.4

1.5

21

24.4

1.16

Average

1.35

1.41

1.90

2.1.3. Wind maximization

Wind maximization is most commonly used in orographic regions. (WMO.332) the wind maximization ratio is simply the ratio is simply the maximum average wind speed for some specific duration and critical direction obtained from a long record of observations, e.g. 50 y, to the observed maximum average wind speed for the same duration in the storm being maximized. Maximum values of wind speed are obtained from maximum persisting 12h. wihd. In order to estimate high wind speed used 50-y return period.

2.2. Statistical Method

Statistical Method for estimating probable maximum precipitation is used wherever sufficient precipitation data are available, and particularly useful for making quick estimate, or where other meteorological data, such as dew point and wind records, are lacking. The statistical method is used mostly for making quick estimate for basins of no more than about 1000 km2, but has been used for much larger areas (WMO, 332). This method requires data for annual maximum precipitation series in the region for required storm durations. The statistical method developed by Hershfield (1961) and modified (1965) is based on the general frequency equation 2 (chow 1961) Xt = Xn+ KSn (2) Where Xt is the precipitation for return period t; Xn and Sn are respectively the mean and standard deviation of a series of n annual maximum: and K is a common statistical variable, which varies with the different frequency distributions fitting extreme value hydrologic data. This method requires some adjustment, such as adjustment Of mean of annual series (Xn), standard deviation of annual series (Sn) for maximum observed precipitation. Such as adjustment of mean (Xn) and standard deviation (Sn) of annual series for maximum observed precipitation and also adjustment of mean and standard deviation of annual series for length of record.

Table (3) estimated probable maximum precipitation value based on statistical (Hershfield) and physically methods
Station Abadan Ahvaz Mahshahr Brojerd Bushehr Dezful Masjed soleiman Omidieh yasoj Statistical (Hershfilds) methods 236 371 362 277 637 418 516 407 356 Physically (maximization of storms) method 92.5 114.7 137.3 98 145.3 151.7 144 93.9 134.9

26 25 24 23 D WP IN (C E O T ) 22 21 20 19 18 17 16 15 1 0 2 0 1 0 2 0 1 0 2 0 1 0 2 0

y = -5E-06x + 0.0004x - 0.0118x + 0.1789x - 1.3036x + 3.2302x + 21.247 2 R = 0.8364

1 0

2 0

1 0

2 0

1 0

2 0

1 0

1 o. -N v

1 e . -F b

1 a . -J n

1 e. -D c

TIME(DAY)

32 31 30 29 28 27 26 D EWP IN ) O T(C 25 24 23 22 21 20 19 18 17 16 15 10 20 10 20 10 20 10 20

y = -8E-06x + 0.0007x - 0.0199x + 0.2907x - 1.9959x + 4.2433x + 27.461 2 R = 0.9711

1 a -M r.

1 p -A r.

1 c -O t.

10

20

10

20

10

20

1 a -M y

10

2 0

1-N ov.

1-Feb.

1-Jan.

1-D ec.

TIME(DAY)

32 31 30 29 28 27 26 DEW POINT(C) 25 24 23 22 21 20 19 18 17 16 15 10 20 10 20 10 20 10 20

y = -4E-06x + 0.0004x - 0.0113x + 0.1667x - 1.0181x + 0.4534x + 31.376 2 R = 0.9162

1-M ar.

1-A pr.

1-O ct.

10

20

10

20

10

20

1-M ay

10
10 20

1-Jan.

1-Feb.

1-Nov.

1-Dec.

TIME(DAY)

29 28 27 26 25 D EW PO T(C) IN 24 23 22 21 20 19 18 17 16 15 10 20 10 20 10 20 10 20 10

y = -4E-06x + 0.0003x - 0.01x + 0.1469x - 0.9727x + 1.23x + 27.867 2 R = 0.9859

1-Mar.

1-Apr.

1-Oct.

20

10

20

10

20

1-N ov.

1-Feb.

1-Jan.

1-Dec.

TIME(DAY)

Figure (3): Maximum dew point envelope curves (a) in the Dezful station

(b) Abadan (c) Ahvaz and (d) Bushehr station

1-M ar.

1-A pr.

1-Oct.

1-M ay

1-May

20

20

Conclusions
In this study two techniques used for estimating PMP which incloude statistical and physically methods. Statistical method based on the transposition and maximization of historical precipitation (annual maximum precipitation series) and physically method based on the maximization of the physical factors (dew point and wind) that the precipitation evolution. Using statistical method of estimating PMP rather than physically method showed larger values. Also application of the physically method will allow for the provision of PMP estimates for catchments in the southwest Of Iran from 1-5000 km2 in size and duration of 24,48,72,96,120 and 144 hours, whereas statistical method used mostly for basins about 1000km2. We found that PMP estimates by statistical method are well comparable with values of obtained by the physical method for different durations. Results also shows that limited transposition of statistical methods gives higher estimates, in comparisons with physical method.

References
Desa, M.N., Noriah,A.b. and Rakhecha,P.R.(2001) Probable maximum precipitation for 24h duration over southest Asian Monsoon region-Selangor Malaysia. www.Elsevier.com/ocate/atmos. Koutsoyiannis, D. (1999)A probabilistic view of Hershfield method estimating probable maximum precipitation. Water resources research.4:1313-1322. Loukas, A and quick, M.C.A. (1996) Spatial and temporal distribution of storm precipitation in southwestern British Columbia; Journal of Hydrology, 174: 37-56 Kennedy, M.R. (1982) The estimation of probable maximum precipitation in Australia- past and Current practice, proc. Workshop on spillway design, Melbourne, 1981. Awrc cont. 26-52 Bureau of Meteorology (1985) ' The estimation of probable maximum precipitation in Australia for short small areas' Bulletin 51, august 1984, AGPS, Canberra. World Meteorological Organization (1986)' Manual for estimation of probable maximum precipitation; second edition. Operational hydrology report No.1, WMO-No.32, and Geneva. Burea of Meteorology (1994)' the estimation of probable maximum precipitation in Australia: Generalized short-duration method. Bulletin 53, December 1994, AGPS, Canberra. Corrigan, p. and J. L. Vogel (1999)' probable maximum precipitation for California- Calculation procedure, Hydro meteorological report No.58.national weather service.