Development of a Digital Recording System for Automation of Class A Evaporation Pan Measurement

Abstract

Adapting a sensor for automation of pan evaporimeters will increase reliability and accuracy of evaporation measurements especially in areas where data reliability is questionable. This research was aimed at developing and calibrating a digital recording (water level sensing device) system for use with Class A pan. A simple quadratic equation (calibration curve) was fitted between the sensor's recorded depth and manually recorded depth (R 2 > 0.98; F = 4388; P < 0.0001). The system performance was evaluated on the basis of its accuracy in measuring water depth when compared with readings from manual observations. Two accuracy parameters, correlation coefficient (R) and mean bias error (MBA) were used to indicate this accuracy. A high R > 0.97 and MBA of 0.19 mm/day was obtained between the digital sensor values and manual readings of pan evaporation for the months of March while R > 0.97 and MBA of 0.15 mm/day was estimated for the month of April. The high correlation coefficient between pan evaporation values from digital sensor and the Manual method is an indication of the accuracy and sensitivity of the developed digital recording system for Class A pan evaporation measurement. The result of this study offers an opportunity for a near real-time evaporation measurement.

Full text

1
J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript 1061. Volume XI. October, 2009.
Development of a Digital Recording System for Automation of Class A
Evaporation Pan Measurement
1Fasinmirin J.T. and 2Oguntunde P.G.
1Department of Agricultural Engineering, Federal University of Technology, Akure, Nigeria.
2Department of Civil Engineering and Geosciences, Technical University of Delft, Netherlands
Corresponding Author’s E-Mail: poguntunde@yahoo.com
ABSTRACT
Adapting a sensor for automation of pan evaporimeters will increase reliability and accuracy of
evaporation measurements especially in areas where data reliability is questionable. This
research was aimed at developing and calibrating a digital recording (water level sensing device)
system for use with Class A pan. A simple quadratic equation (calibration curve) was fitted
between the sensor’s recorded depth and manually recorded depth (R2 > 0.98; F = 4388; P <
0.0001). The system performance was evaluated on the basis of its accuracy in measuring water
depth when compared with readings from manual observations. Two accuracy parameters,
correlation coefficient (R) and mean bias error (MBA) were used to indicate this accuracy. A high
R > 0.97 and MBA of 0.19 mm/day was obtained between the digital sensor values and manual
readings of pan evaporation for the months of March while R > 0.97 and MBA of 0.15 mm/day
was estimated for the month of April. The high correlation coefficient between pan evaporation
values from digital sensor and the Manual method is an indication of the accuracy and sensitivity
of the developed digital recording system for Class A pan evaporation measurement. The result
of this study offers an opportunity for a near real-time evaporation measurement.
Keywords: Pan evaporation, automation, irrigation scheduling, sensor, climate, Nigeria
1. INTRODUCTION
Accurate measurements of environmental variables such as rainfall, runoff, soil moisture,
evaporation rates, minimum and maximum temperatures, etc., are essential aspect of
agrometeorologial research that is important in agricultural related issues. Evaporation is an
important climate element as it affects both plant and animal life and is a major factor in man's
comfort and well-being (Purvis, 2006). Design of irrigation systems and scheduling of irrigation
requires information on evaporation rates to the atmosphere as well as transpiration from plants.
Evaporation pans can be used to indicate the rate of crop water use (Smajstrla et al., 2000). Data
from evaporation pans are a great tool to study climate and this has been done by several
researchers (Brutsaert and Parlange, 1998; Linacre, 2004; Liu et al., 2004; Roderick and
Farquhar, 2004) Pan Evaporation is a measurement that combines or integrates the effects of
several climate elements: temperature, humidity, solar radiation, and wind (Allen et al., 1998;
Al-Ghobari, 2000).
Of the many avenues for water loss, evaporation remains the most important in most part of
Nigeria. However, routine measurement of evaporation in Nigeria still suffers from poor
coverage and non-uniform instrumentations (Owonubi and Olorunju, 1991). Some weather
stations include measurements of the evaporation from a US Class-A pan as recommended by
the Federal Department of Meteorology in Nigeria. But most places are without such
measurements, and, even where there is a pan, the measurements may be smeared by poor

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
maintenance, leading to errors due to leaks, the growth of algae in the water, an incorrect water
level, weed-growth nearby, un-reliable attendants, and so on.
Recently, precision agriculture technologies have made significant advances in the area of
irrigation scheduling, especially in developed countries, where equipment for continuous
monitoring of climatic conditions is now available to help producers determine how much water
to apply and when to apply it. Phene (1992) and Phene (1995) showed that frequent
measurement of evaporation rates from an automated Class A evaporation pan can accurately
estimate ET and be used as an irrigation scheduling tool. Key to Phene's work is the finding that
ET data is more accurate when the water level in the pan is measured hourly making it possible
for real-time irrigation scheduling of crops with an automated pan evaporation system.
However, Investments in such espensive automated system is more risky in the cash-strained
economic realities of African farmers. Therefore, the use of electronic level sensors for
automation of pan evaporimeter has been given little or no attention in this region. This project
stems from the need for a more accurate method of measuring evaporation as well as better
estimate of crop water use. Availability of automated system will minimize the incidence of
human error in the measurement of evaporation from the pan evaporimeter and provide avenue
for real-time irrigation scheduling. Therefore, the objective of this study was to develop and
calibrate a cheap electronic device for water level measurement for use with Class A pan.
2. MATERIAL AND METHODS
2.1 Description of class A pan
A Class A evaporation pan is cylindrical with diameter of approximately 121 cm and 25
cm deep. Usually was made of twenty-two (#22) gauged galvanized metal sheet and mounted on
an open frame with its bottom 15 cm above the ground. A typical Class A pan evaporimeter is
shown in Figure 1. The pan rests on a carefully leveled, wooden base and is often enclosed by a
chain link fence to prevent animals drinking from it. Evaporation is measured daily as the depth
of water (in mm) evaporates from the pan. The measurement day begins with the pan filled to
about 5 cm from the pan top. At the end of 24 hours, enough water is added, in measured
increments, to again fill the pan to initial level from its top. If precipitation occurs in the 24-hour
period, it is taken into account in calculating the evaporation. Sometimes precipitation is greater
than evaporation, and measured increments of water must be dipped from the pan.
Figure 1. A typical class A evaporation pan (www.novalynx.com)

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
2.2 General description of the digital sensor
The direct display drive (ICL7106) is a high performance, low power 3½ digit analog to digital
converter. All the necessary active devices are contained on a single CMOS, including seven
segment decoders, display drivers, reference and clock. The ICL7106 was designed to interface
with a liquid crystal display (LCD) and include a back plane drive (Cox, 1988).
The digital device, ICL7106 brings together a combination of high accuracy, versatility and low-
cost; high accuracy like auto-zero to less than 10μv, zero drift to less than 1Nu/c, input bias
current of 10 pA maximum, and roll over of less than one count. The versatility of true
differential input and reference is useful and advantageous when measuring load cells, strain
gauges and other bridge-type transducers. The true economy of a single power operation (7106)
enables a high performance panel meter to be built with the addition of only seven (7) passive
components and a display. The resistance (R) varies according to the level of water in the pan. It
divides the reference voltage (2v) in ratio to the value of the internal resistance (Ri). This varying
voltage is now converted to digital readout by the A/D converter. The schematic representation
of the ICL7106 is shown in Figure 2.
Figure 2: The ICL7106 with liquid crystal display
V+
1
D1
2
C1
3
B1
4
A1
5
F1
6
G1
7
E1
8
D2
9
C2
10
B2
11
A2
12
F2
13
E2
14
D3
15
B3
16
F3
17
E3
18
AB4
19
POL
20
OSC 1
OSC 2
TEST
REF HI
REF LOW
C
REF+
C
REF-
COM
IN HI
32
IN LO
31
A-Z
30
BUFF
29
INT
28
V-
27
G2
26
C3
25
A3
24
G3
23
GND
22
OSC 3
21
40
39
38
37
36
35
34
33
TO DECIMAL POINT
-5 V
+5 V
R3/100 k
C4/100 p
VR6/10 k
R1/18 k
C1/0.1 µ
R4/1 M
C5/0.01 µ
C2/0.47 µ
C3/0.22 µ
R2/470 k
R5/100 R

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
An internal digital ground is generated from a 6v zener diode and a large P-channel source
follower in the digital section of the ICL7106. This supply is made stiff to absorb the relative
large capacity currents where the back plane (BP) voltage is switched. The BP frequency is the
clock frequency divided by 800 dings per second (Cox, 1991), this s a 60Hz square wave with
nominal amplitude of 5v. The segments are drives at the same frequency and amplitude are in
phase with BP when “OFF”, but out of phase when “ON”. In all cases negligible DC voltage
exists across the segments.
A circuit was built, as shown in Figure 3, and connected to pin 38 of the ICL7106 in order to
generate a stable reference voltage for the divider (sensor). The entire system was powered by
two 9-volt batteries. The first battery provided power to the circuit while the second powered the
LCD and provided the reference voltage. The sensor being a resistive-type acted as a voltage
divider. Initially, the pan is at zero level and the entire system is having a voltage of 2 volts
across it, which is the reference voltage, and the output at the display end is 00.0. As water level
in the pan decreases due to evaporation, the arm of the sensing device moves and this movement
accentuates a slider moving across a coil of thin wire. The resistance in the coil varies according
to the position of the slider and thus the movement of the slider causes a change in resistance,
which in turn causes the voltage across the system to change. The ICL7106 being an A/D
(Analog to Digital) converter then converts the new voltage from electrical impulses to a digital
display.
Figure 3: The built and connected circuit
2.3 Calibration procedures for the digital sensor
The calibration procedure is as follows:

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
i. The pan was filled to its maximum allowable level (zero level) where the reading on the
LCD was 0.00
ii. The water level in the pan was gradually reduced by 1 mm pan depth for every 2.5cm3 of
water removed from the pan and the LCD output taken at each 1 mm depth.
iii. The results obtained were plotted on a graph and a calibration curve fitted between
manual and sensor readings of water depth in the pan.
2.4 Performance evaluation
The performance of the developed digital sensor was evaluated to ascertain its accuracy and
sensitivity in measuring pan evaporation (Ep). The digital values obtained were first converted
using the calibration equation and then compared with the Ep values obtained from manual
measurements for the months of March and April 2005. Statistical analysis, to determine the
degree of associations between data obtained from sensor and manual readings are correlation
coefficient (R) and mean bias error (MBE). These were used to indicate the sensor accuracy.
MBE, which quantifies the bias of the measurements, was computed as:
*
1
1()
n
ii
i
MBE Ep Ep
n
=
= −
∑
(1)
where Epi* is the sensor value and Epi is the manual value. MBE should be close to zero for
unbiased sensor reading, while the value of R explains the degree of associations between the
measurements. Correlation R ranges from 0 to 1 and the closer to unity the better.
3. RESULTS AND DISCUSSION
Calibration curves based on linear regression equation are shown in Figure 5A while Figure 5B
showed calibration curve based on a quadratic model. A summary of the calibration statistics is
presented in Table 1. Results showed that the sensor readings are almost linear with the manual
observation with R2 > 0.96 in both cases. The quadratic calibration curve showed a better fit with
R2 > 0.98. All the models are highly significant at P < 0.0001. Although both linear fits showed
high accuracy with their R2 not statistically different from that of the quadratic model, the values
of standard error (0,111 mm and 0,150 mm) are higher making them inferior to the non-linear
equation which has SE of 0.079 mm. Linear model starting from the origin is theoretically
preferred over the other since no negative water depth was observed. However, this equation will
lead to high errors especially at water depth below 2.5 mm as revealed by the scatter points away
from the trend line of Figure 5A (dashed line). Given the foregoing, calibration curve based of
the quadratic model was adjudged best and then adopted for this electronic water level sensor.
Table 1: Summary of calibration statistics
S/N
Model
#
Overall Statistics
R2
Standard error
Sig. level
1
Linear with intercept
0,9796
0,111
< 0.0001
2
Linear without intercept
0,9613
0,150
< 0.0001
3
Quadratic
0,9897
0,079
< 0.0001
#R2 is coefficient of determination; SE is standard error of estimates
The performance was further validated by measuring Ep simultaneously with the sensor and
manually for the months of March and April. Ep obtained from digital sensor conversion using

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
the quadratic calibration equation were very similar to values from manual measurement as
shown Figures 5 and 6, for March and April, respectively. The variation between both the
manual and digital datasets was quite small. MBE for the month of March and April were 0.19
mm/day and 0.15 mm/day, respectively and the correlation coefficients were very high (R >
0.97) for both datasets.
A
y = 0,5745x - 0,2259
R
2
= 0,9796
y = 0,5043x
R
2
= 0,9613
-0,5
0,0
0,5
1,0
1,5
2,0
2,5
3,0
0,0 1,0 2,0 3,0 4,0 5,0
Water depth (Sensor, mm)
Water depth (manual, mm)
B
y = 0,0512x
2
+ 0,3235x
R
2
= 0,9897
0,0
0,5
1,0
1,5
2,0
2,5
3,0
0,0 1,0 2,0 3,0 4,0 5,0
Water depth (Sensor, mm)
Water depth (manual, mm)
Figure 4: Calibration curves for digital sensor (A) linear fits with intercept (solid line) and
without intercept (dashed line); and (B) quadratic fit without intercept.

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
0,0
1,0
2,0
3,0
4,0
5,0
6,0
0 5 10 15 20 25 30 35
Day of the Month (March)
Ep (mm/day)
Manual Sensor
Figure 5. Pan evaporation (Ep) measured with digital sensor and manually in the month of March
2005.
0,0
1,0
2,0
3,0
4,0
5,0
6,0
7,0
0 5 10 15 20 25 30
Day of the Month (April)
Ep (mm/day)
Manual Sensor
Figure 6. Pan evaporation (Ep) measured with digital sensor and manually in the month of April
2005.
4. CONCLUSION
A digital electronic device was developed and calibrated for use in evaporation measurement
with Class A pan evaporimeter. A quadratic calibration curve of the form:
2
ss
0.0512Ep 0.3235Ep 0.079Ep = +±
, R2 = 0.989 (2)

J.T. Fasinmrin and P.G. Oguntunde. “Development of a Digital Recording System for
Automation of Class A Evaporation Pan Measurement”. Agricultural Engineering International:
the CIGR Ejournal. Manuscript LW 1061, Vol XI. October 2009.
was fitted for the digital sensor. Where Ep is the actual pan evaporation and Eps is the sensor
reading. Observations from two months of experimentation indicated that the digital device has a
high degree of accuracy, easy to use and reduces the time spent on field when compared with
manual method. The digital water level sensor presented offers opportunity for near real-time
measurement of evaporation and/or possibility of remote logging. However, improvement is
required especially to incorporate an automatic re-fill mechanism to the present design.
5. REFERENCES
Al-Ghobari, H. M. 2000. Estimation of reference evapotranspiration for southern region of Saudi
Arabia, Irrigation Science 19: 81-86.
Allen, R.G., L.S. Pereira, D. Raes and M. Smith. 1998. Crop Evapotranspiration: Guidelines for
Computing Crop Water Requirement. Agron. 18: 289 – 307. Irrig. And Drain paper
56 FAO, Rome.
Brutsaert, W., and M. B. Parlange.1998. Hydrologic cycle explains the evaporation paradox,
Nature, 396, 30.
Cox, S.W.R. 1988. Farm Electronics, BSP Professional Books. A division of Blackwell Scientific
Publications Ltd. Beacon Street, Boston Massachesetts 02108, USA. pp 171-174.
Linacre, E.T. 2004. Evaporation Trends. Theor. Appl. Climatol. 79: 11-21.
Liu, B. M. Xu, M. Henderson, and W. Gong . 2004. A spatial analysis of pan evaporation trends
in China, 1955–2000, Journal of Geophysical Research, Vol. 109, D15102,
doi:10.1029/2004JD004511.
Owonubi, J. J. and Olorunju, S. A. S. 1991. A comparative study of evaporation rates measured
by different instruments at Samaru. Samaru J. Agric. Res., 8: 75-80.
Phene, C.J.,W.R. DeTar, and D.A. Clark. 1992. Real-time irrigation scheduling of cotton with an
automated pan evaporation system. App. Engineer in Agric, 8(6):787-793.
Phene, R.C. 1995. Class "A" evaporation pan for irrigation scheduling in real-time. Proc. Inter.
Expo. And Tech. Conf., Phoenix, AZ. p153-160.
Purvis, J. C. 2006 Pan Evaporation Records for the South Carolina Area, State Climatological
office. http://water.dnr.state.sc.us/water/ asseced on 17-11-2006.
Roderick M. L. and Farquhar G. D. 2004. Changes in Australian Pan Evaporation from 1970 to
2002, Int. J. Climatol. 24: 1077–1090.
Smajstrla A.G., F. S. Zazueta, G. A. Clark, and D. J. Pitts. 2000. Irrigation Scheduling with
Evaporation Pans. Bulletin 254, the Department of Agricultural and Biological
Engineering, Florida Cooperative Extension Service, Institute of Food and
Agricultural Sciences, University of Florida. http://edis.ifas.ufl.edu.