Archive-name: meteorology/temp-dewpoint
Posting-Frequency: About monthly.
Version: 004
Date: September 19, 1996
Updated: When necessary.
Lines: about 550.


  The _Temp, Humidity & Dew Point_  ONA (Often Needed Answers)




Table-of-contents:
1)  Introduction.
2)  Formulae.
3)  Examples.
4)  Literature.
5)  Committment.
6)  Outlook.
7)  Signature.



1)
Introduction:

From the discussions on the newsgroup sci.geo.meteorology this is a
collection of some formulae and texts that reflect on connections
of temperature, humidity and dew point temperature (BeK):

Air will normally contain a certain amount of water vapour. The
maximum amount of water vapour, that air can contain, depends on
the temperature and, for certain temperature ranges, also on whether
the air is near to a water or ice surface. If you have a closed con-
tainer with water and air (like a beaker) then there an equilibrium
will develop, where the air will contain as much vapour as it can.
The air will then be saturated with respect to water vapour.
The real world outside is not closed, so that the air normally will
contain less vapour as it could. Sources of vapour are evaporation
processes from water and ice surfaces and transpiration from plants
and respiration from animals. The expression "evapotranspiration"
takes into consideration plants' large share of evaporation over
land areas.
Sinks of water vapour are clouds or condensation on surfaces.
Dew is created when a surface temperature has such a low temperature
that the air chills to the dew point and the water vapour condenses.
Physically at the dew point temperature the vapour loses the energy
that it gained at evaporation, the latent energy, again.

The precipitable water (total column water vapor) is strongly
correlated (r > 0.9) with the surface dew point on most days.
Exceptions to the rule include days when a cold front has passed
and during other transient events. (Kerry Andersen)



2)
Formulae:

Enough for dry physical theories; here comes the practice.

For some people skipping this and going directly to the examples
would be the most rewarding. Especially as they treat the conversion
of relative humidity and psychrometer temperatures. (BeK)

Vapor pressure (e) is the fraction of the ambient pressure that is
due to the fraction of water vapor in the air.
Saturation vapor pressure (es) is the maximum vapor pressure that the
air can support (non supersaturated) at a given temperature.

e  can vary from 0 (verrry dry) to the maximum, es.
es is a function of temperature es(T).


Relative humidity (RH) is 100% times the ratio of the environmental
vapour pressure, e(T), to the saturation vapour pressure es(T).

        RH = 100% * e(T)/es(T)


The environmental vapour pressure is the saturation vapour pressure
at the dew point or

        e(T) = es(Td)

so RH becomes

        RH = 100% * es(Td)/es(T)

In other words: if you have a parcel of air and cool it until the
water vapor in it condenses then you have reached the saturation point.
At this point you will measure the same vapour pressure as in your
original air probe.



Some more elaborate expressions follow here:

es0 = reference saturation vapor pressure (es at a certain temp,
                                           usually 0 deg C)
    = 6.11 hPa

T0  = reference temperature (273.15 Kelvin,  Kelvin = degree C +
                                                      273.15)

Td  = dew point temperature (Kelvin)

T   = temperature (Kelvin)

lv  = latent heat of vaporization of water (2.5 * 10^6 joules
                                                       per kilogram)

Rv  = gas constant for water vapor (461.5 joules* Kelvin / kilogram)

e  = es0 * exp( lv/Rv * (1/T0 - 1/Td))
es = es0 * exp( lv/Rv * (1/T0 - 1/T))

e/es * (100%) = relative humidity !!!!

So just above is the answer to many questions in the direction of
how to calculate the relative humidity if you have the dew point
and air temperature.



There are some simple and more complicated formulas for the
saturation vapour pressure at a given temperature.

A simple first guess (assuming the latent heat of vaporization is
constant with temperature) would be:

        log10(es) = 9.4041 - 2354/T

where T is in Kelvin (i.e., 273.16+T(C))

Another approximation (Magnus' formula) would be

        log10(es) = -2937.4/T - 4.9283*log10(T) + 23.5470



In the following the input gets a little more complicated. Here we
also shall distinguish between the saturation vapour pressure over
ice or water. Both are different, as the molecular forces bind much
more in an ice crystal than in a water bobble. So the saturation
pressure esW will be larger than esI (W for water, I for ice).

 1.  Vapor pressure (ED):

                             dew point temperature in degrees C.
                            /
     ED = 6.1078 * 10 ** ((TD * A)/(TD + B)) in mbs


 2.  Saturated vapor pressure (ES):


     ES = 6.1078 * 10 ** ((T * A)/(T + B))
                            \
                             temperature in C

                   A =   7.5 } for use in
                   B = 237.3 } vapor pressure with respect to water

                 * A =   9.5 } for use in vapor pressure
                   B = 265.5 } with respect to ice



 3.  Absolute virtual temperature (TV):

                                 vapor pressure
                                /
     TV = (T + 273.15)/(1-0.379*ED/Press)
                                    \
                                     total pressure

     TV does take into consideration that you could try to condense
     all the water vapour in your air parcel and use the condensation
     heat to warm up the air. This is a first way to distinguish
     different air parcels that may have the same temperature but
     have different relative humidity.



 4.  Mixing ratio (W):
                                 vapor pressure
                                /
                        .62197 e                      grams water
                    W = --------                      -------------
                           P - e                      grams dry air
                            \                              |
                             total pressure                |
                                                   Thus 12 g/Kg comes
out
                                                   as .012


 5.  Wet Bulb
     Vapor Pressure
     Dew Point               (P365, Smithsonian for first part)

     Ew - E
     ------      = .000660 (1 + .00115 T )
     Press (T-Tw)                       w

 Therefore:

            E = Ew - Press (T-T ) (.000660) (1 + .00115 T )
                               w                        w
           Tw = Wet bulb temperature (degrees C.)
           Ew = Saturated vapor pressure at temperature Tw
            E = Vapor pressure in air
        Press = Total barometric pressure (units same as Ew, E)
            T = Air temperature (degrees C.)

 E is the vapor pressure in the air, which is the vapor pressure at
 the dew point temperature.  To solve for the dew point temperature,
 use the formula:

            E = 6.1078 * 10 ** ((TD * A)/(TD + B)) in mbs

                 let C = log   (E/6.1078)
                            10

 Then:

            C T  + C B = A T
               d            d

                    B*C
               T =  ---     Dew point in degrees C
                d   A-C
                                           where A =   7.5
                                                 B = 237.3




All the above saturation pressure temperature relationships are
relatively uncomplicated. Here one that is more mindboggling:

A saturation-pressure-curve which is valid for a total pressure of
1000 hPa. "This curve was computed by approximating the standard
steam table for pure water using the least square method by a
Bulgarian colleague. I experienced it to be quite exact, but I'd
be glad to be corrected." (Dr Haessler)

Psat  = 610.710701 + 44.4293573*t + 1.41696846*t^2 +
        0.0274759545*t^3 + 2.61145937E-4*t^4 + 2.85993708E-6*t^5

The pressure is in Pa, the temperature in degrees Celsius (C).


Relative Humidity then is:

Phi = Psteam/Psat = (Ptot/Psat)*x / ((Rair/Rsteam)+x),

where x is the absolute humidity in kilogramm water per kilogramm
of dry air,
Rair and Rsteam are the specific gas constants for air and steam,
where
Rair/Rsteam has a value of 0.622.

For the handling:
1. Calculate the saturation pressure at Your dew point, giving Your
   steam pressure.
2. Calculate the saturation pressure at Your temperature.
3. Divide'em (see above) to get Your relative humidity.
4. Calculate Your absolute humidity, if desired.
5. Mail me for further informations, if necessary.
6. The reverse way is possible.

For the pressure dependence of relative humidity:
If air and steam behave as ideal gases, there is no pressure
dependence.
This is so around 1000 hPa (+-100hPa, approx.).



3)
Examples

1. EXAMPLE      X M P L    X M P L    X M P L    X M P L    X M P L
>My problem is the following. I want to calculate wetbulb temperature
>(Tw) where my input is drybulb temperature (TT) and relative humidity
>(rH). (Pieter Haasbroek)

Pieter:

Your problem can be solved explicitly using the methods from Jensen et
al. (1990) ASCE Manual No. 70 (see pages 176 & 177) using the
following steps and equations:

1)  compute ED as [EA(TT)*rH/100]
    where EA(TT) = 0.611*EXP(17.27*TT/(TT+273.16)) in kPa
    TT is drybulb temp in C

    ED = (rH/100)*0.611*EXP(17.27*TT/(TT+373.16))
    where ED is ambient vapor pressure in kPa

2)  compute dewpoint temperature (Td)
    Td = [116.9+273.3ln(ED)]/[16.78-ln(ED)] in C

3)  compute wet bulb temperature (Tw)
    Tw = [(GAMMA*T)+(DELTA*Td)]/(GAMMA+DELTA)
    GAMMA = 0.00066*P where P is ambient barometric pressure in kPa
    DELTA = 4098*ED/(Td+237.3)^2

This method should be close, especially when Tw is close to Td (DELTA
should be evaluated at (Tw+Td)/2.

For example:

TT = 25C
RH = 50%
assume elev is sea level and P = 100 kPa.

1)  EA(25) = 0.611*EXP(17.27*25/(25+273.16)) = 3.17 kPa
    ED = (50/100)*2.60 = 1.59 kPa

2)  Td = [116.9+237.2*ln(1.59)]/[16.78-ln(1.59)] = 13.9C

3)  GAMMA = 0.00066*100 = 0.066 kPa/C
    DELAT = 4098*(1.59)/(13.9+237.3) = 0.1033 kPa/C
    Tw = [(0.066*25)+(0.1033*13.9)]/(0.066+0.1033) = 18.2

CHECK ANSWER:

    EW(Tw) = 0.611*EXP(17.27*18.2/(18.2+237.16)) = 2.09 kPa
    ED = EW(Tw) - GAMMA*(T-Tw)
    ED = 2.09 - 0.066*(25-18.2) = 1.64 kPa

    ERROR ED = [(1.64 - 1.59)/1.59]*100 = 3.1%
    ERROR Tw = [(18.2-17.95)/17.95]*100 = 1.4%

    Exact answer for Tw is about 17.95C
    EW(18.0) = 2.07 kPa; ED = 1.60 kPa
    EW(17.9) = 2.05 kPa; ED = 1.58 kPa
    EW(17.95) = 2.06 kPa; ED = 1.59 kPa

CAUTIONS:

Wet bulb temperature is really defined by the psychrometer and is not
an atmospheric water vapor property (compared with Td which has a firm
definition)!!!!   The above computations assume the standard
psychrometer equation, but the psychrometer constant (0.00066*P in
kPa/C) is a theoretical value that is not always matched even by very
good psycrchrometers.  PLEASE note this psychrometer constant depends
directly on atmospheric pressure so it's value is not a "universal"
constant!

Terry Howell



2. EXAMPLE      X M P L    X M P L    X M P L    X M P L
>Hello:-
>I am looking for the algorithm to convert wet/dry bulb temperatures to /
>from RH (and moisture content as well, for that matter).
>I know the Psychometric charts, but they are difficult to use accurately
>in software. Anyone have a pointer to appropriate equations?
>Thanks in advance! (Spehro Pefhany)


Answer
(I shall find a formula in SI units, please be patient, BeK)

pw = psf  - p * A * (theta - thetaf)

theta: dry bulb temp., Kelvin or Celsius
thetaf: wet bulb temp.,   "
psf: Saturation pressure at temp thetaf, see 1.), in Torr (mm Hg)
pw: Vapor pressure of ambient air, in Torr (mm Hg)
p: pressure of ambient air, in Torr
A: optimally (see below) 0.66 * 10e-3 * (1/C)

3.) The short way round:

We're in your backyard: p = 755 Torr, 0 C < theta < 50 C.

pw = psf - k (theta - thetaf)

phi = pw/psf

phi: relative humidity.
k= p*A = 0.5 Torr/degree

Higher temperatures:
   thetaf about 60 C: k is about 0.52
   thetaf about 80 C: k is about 0.53


(Formula suggested by A. Sprung, 1888)



Good psychrometers

a) Air velocity

k (and A) don't really become (sorta) independent of the air velocity
past your wet bulb until velocities above 3 meters/ second.
Velocities greater than 1 m/s are sufficient at temperatures of 60 C
or more.
The worse your arrangement, (less adiabatic, i.e. the more extraneous
energy radiates/conducts into the water) the steeper k over velocity
becomes for lower velocities. So you can compensate poor design to
some extent by cranking up that fan.

k and A are really device-dependent. This k (and A, of course)
strictly refers to the "Assmann psychrometer" only - two radiation
shields, thermal insulation, fan downstream from the thermometers. k
should be similar for any well-made psychrometer. (The guy is German,
and dead anyway, so his name can't really bother him anymore...).

b) Adiabatic wet bulb

Shield it from radiative & conductive errors, i.e. all energy to
vaporize the water must come from the air and thus be reflected in
thetaf.

In wetting the wet bulb, use distilled water. Salty scale on your
"sock" can change the vapor pressure, and will really mess
measurements near zero. Use enough water to hit steady-state
conditions well before you start to dry out.

If you use a wick for continuous wetting, make it long enough so that
conductive errors are minimized, and it is cooled to the wet bulb
temperature by the time it gets near the thermometer. Make sure enough
water can reach the wet bulb, so don't overdo the "long enough" part.

Don't get anything but the wet bulb wet. Getting the radiation shield
or the thermal insulation wet will introduce errors.

Keep direct sunlight off. A great way to pump heat into your
"adiabatic" system. Don't ever paint the outside black. Many
commercial humidity meters are a pretty black finish. They will be
sensitive to indirect sunlight (and other radiative sources). Humidity
measurements are VERY sensitive to temperature!

c) Supercooled water and ice below freezing

Your measurement will become screwy below freezing, as you cannot
really distinguish between supercooled water (evaporation) and ice
(sublimation) in your wet bulb, and the vapor pressures differ. And
Lueck says supercooled water can be present as low as -12 C. It
suggests manually scraping the wet bulb to ensure that supercooled
water turns to ice.

And note that humidity measurements never are terribly accurate, 2%
error in absolute hum. are pretty good, depending on where you are in
terms of temp and water content. Anything that reads "relative
humidity=52.783 %" is guessing (if you paid less than 100k$...:-)

Thomas Prufer



4)
Literature hints:

For the book and paper aficionados of the readers check out this:

'If your really interested in this stuff, I (Kerry Anderson) suggest
the book "Atmospheric Thermodynamics" by Irabarne and Godson."'
But unfortunately I learned this book is out of stock
(amazon.com). Instead I could recommend:
"Fundamentals of Atmospheric Dynamics and Thermodynamics"
Paperback, Amazon.com Price: $29.00; Published by
World Scientific Pub Co. Publication date: May 1992
ISBN: 9971978873

"Most introductory texts on meteorology will have one
 or two paragraphs on the matter." (K Anderson)



"My sources (other than experience) are all German books
(Thomas Prufer) (ue equals u&uml;, BeK):"

Lueck, Winfired: Feuchtigkeit - Grundlagen, Messen, Regeln.
Muenchen: R. Oldenbourg, 1964. Good basics.

Sonntag, D.: Hygrometrie: Ein Handbuch der Feuchtigkeitsmessung in
Luft und anderen Gasen. (6 vols.) Berlin: Akademie, 1966 - 1968
Also contains a very detailed description of nearly everything on the
market in 1966-68.

Heinze, D.: Einheitliche, methodische Beschreibung von
Gasfeuchte-Messverfahren. Dissertation an der Technischen Hochschule
Ilmenau, 1980
Comprehensive block and signal diagrams with the Laplace functions (!)
of nearly all humidity measurement methods. Nearly unobtainable,
unfortunately.
(Thomas Prufer)




5)
Committment:
This ONA was collected and provided to you by Bernd Kuemmel
(bek@mmf.ruc.dk).

I admit to have used especially the willing help and the contributions
of the of the following people:

Forrest M. Mims III, Kerry Anderson, Len Padilla, Ralf Haessler,
Pieter Haasbroek, Terry Howell, David F Palmer, Thomas Prufer,
Spehro Pefhany, and of course Ilana Stern

during the ongoing improvement of the ONA.


                        Yours sincerely

                         Bernd Kuemmel



6)
Outlook:

Apparently the psychrometers are very sensible instruments and
errors can sneak in easily. This point I shall elaborate on in
the next versions, BeK.

A plain text version of this text can also be found on:
http://mmf.ruc.dk/~bek/relhum.htm


7)
Signature:
       Bernd Kuemmel + bek@mmf.ruc.dk + VOX: +45 46 75 77 81 * 2275
       IMFUFA, Roskilde University Centre, PB 260, DK-4000 Roskilde
         Disclaimer: They do not necessarily agree with all this.


