Two types of stations are being used to monitor the components of the energy budget. One type is used at locations where plant transpiration could be an important part of the total ET. This type of station features a pair of movable air temperature and humidity sensors for measuring DT and De. These sensors are mounted on an exchange mechanism so that the sensor positions are reversed every 15 minutes. This reversal of position makes it possible to eliminate the effect of sensor bias on the difference measurements and is necessary to measure accurately the small differences in temperature and humidity that occur over the 3 to 5-ft vertical distance. The effect of sensor bias is removed from the 30-minute average by simply averaging the differences measured during two successive 15-minute intervals.
The other type of monitoring station is much simpler in operation and is used where permanent areas of open water occur with little emergent vegetation, so that plant transpiration is an insignificant part of the total ET. At such open-water stations, the same Bowen-ratio principle is used to partition the energy flux into convective and evaporative components. However, the temperature and vapor pressure differences can be measured from water to air, rather than within the air. Since the water-to-air differences are much greater than differences in the air over similar distances, the effect of air and vapor pressure sensor bias is insignificant. Therefore, the sensor exchange mechanism is not required, and only one vapor pressure sensor is needed. Vapor pressure at the water surface can be calculated from the water temperature.
Data are recorded at 15-minute intervals at all of the stations used in vegetated areas, and at 30-minute intervals at the open-water stations. Some of the types of data recorded at the ET stations are not used in determination of ET but are needed in models of ET, or in general interpretation of the data. Multiple sensors are used for soil data because of the heterogeneous nature of soils, and the average of the soil heat fluxes is used in the energy balance. The types of data recorded include rainfall, wind velocity, incoming solar radiation, water level, soil temperature, soil heat flux, water temperature, air temperature, moisture content of air, and moisture content of soil.
There probably are still some more subtle problems with the temperature gradient and vapor-pressure gradient data. These gradients are small in magnitude and may be perturbed at times by insect or moisture contact with the sensors. These types of perturbations are relatively infrequent and probably have little effect on the daily ET sum, though some 15-minute values of computed ET could be off.
A summary of the energy-budget method of ET calculation may help to describe usage of the data files. The basic energy-flow equation is:
Q + G + W + Le + H = 0, (eq. 1)
where Q is net solar radiation, G is soil heat flux, W is heat storage in water (above land surface), Le is the latent heat (ET), and H is the sensible heat (convection). Values for the "W" term are not included in the data files supplied, but are a function of water depth, heat capacity of water, and the mean change in water temperature from one 15-min interval to the next.
The sign convention used with these data is that energy flow is "+" for energy into the ET unit, and "-" for energy out of the unit. Thus, net radiation (Q) is "+" during the day, and "-" at night, and W is "-" when the surface water is heating up during the day and "+" when the water cools. Latent heat (Le), or ET, is "-" while ET is occurring and (theoretically) "+" while condensation, or dew, is forming although dew formation may not actually be able to be measured with the equipment.
The following is a listing of a simple SAS routine used for preliminary ET calculation. It should illustrate most of the details of using the data in the above-listed files. In this routine, the vapor-pressure gradients (de) are automatically screened to exclude values < -0.05 Kpa or > 0.05 Kpa. When values outside this range occur, the last value within the acceptable range is used. This is a rather crude data screening procedure, and other ways for filtering out questionable data are being evaluated.
data atemp; infile 'AIR.TEMP';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 atemp;
run;
data de; infile 'VAP.PRES.GRAD';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 de;
if de < -9998 then delete;
run;
data dt; infile 'AIR.TEMP.GRAD';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 dt;
if dt < -9998 then delete;
run;
data q; infile 'NET.RAD';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 q;
run;
data g ; infile 'SOIL.HEAT.FLUX';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 g;
run;
data wts ; infile 'TOP.WAT.TEMP';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 wts;
run;
data wtb ; infile 'BOT.WAT.TEMP';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 wtb;
run;
data stage ; infile 'WATER.LEVEL';
input @1 date yymmdd8. hr 10-11 min 13-14 @15 stage;
run;
data all; merge atemp de dt q g wts wtb stage; by date hr min;
gamma = 0.000646*101.3*(1.0+ 0.000946*atemp);
avewt = mean (of wts wtb);
run;
data fin; set all;
y=year(date); m=month(date);
lgamma = lag(gamma);
lavewt = lag(avewt); delwt = avewt-lavewt;
depth = stage-5.2;
wheat = -delwt*depth*0.3048*1000000/(14.34*15);
avegam = mean (of gamma lgamma);
if de < -50 | de > 50 then de = lag(de);
bowen = avegam*dt/de;
if bowen > -1.4 & bowen <= -1. then bowen = -1.4;
if bowen < -0.6 & bowen >= -1. then bowen = -0.6;
avail = q + g + wheat;
latent = -avail/(bowen+1.); sense = bowen*latent;
lamda = 2500.25-2.365*atemp;
et = -0.002362*15*latent/lamda;
run;
proc means sum; by y; var et; run;