Content uploaded by Spencer Riley
Author content
All content in this area was uploaded by Spencer Riley on Nov 19, 2019
Content may be subject to copyright.
ATMOSPHERIC PRECIPITABLE WATER AND ITS CORRELATION WITH
CLEAR SKY INFRARED TEMPERATURE READINGS: DATA ANALYSIS
SPENCER RILEY1, VICKI KELSEY2
1DEPARTMENT OF PHYSICS, NEW MEXICO TECH SOCORRO NM, 2LANGMUIR LABORATORY FOR ATMOSPHERIC RESEARCH
REFERENCES
[1] A. Berk, L. S. Bernstein, and D. C. Robertson. MOD-
TRAN: A moderate resolution model for LOWTRAN.
Technical report, July 1987.
[2] Vicki Kelsey. Atmospheric precipitable water and its cor-
relations with clear sky infrared temperature readings:
field observations. Poster presented at PhysCon, Provi-
dence, RI., 2019.
INTRODUCTION
The goal of this research is to determine the correlation
between precipitable water and zenith sky temperature
using low-cost instrumentation.
The methods of measuring precipitable water in-
clude [2]:
•Radiosondes
•Analyzing signal delay from ground-based
Global Positioning System networks.
•Microwave-Infrared radiometers
We can define the relationship between the precipitable
water (TPW) and the measured effective temperature
(Teff) as,
dTeff
dTPW =∂Teff
∂TPW +∂Teff
∂Tair
∂Tair
∂TPW .(1)
Existing models show that there are equal contributions
from both terms in Equation (1), where we define Tair
as the mean temperature within the lowest 5 km of the
atmosphere.
Figure 1: Based on calculations using MODTRAN [1]. Blue, orange, and green
curves relate to TPW of 11.4 mm, 22.7 mm, and 45.4 mm, respectively. Cor-
responding circles represent mean radiance of the spectral band between 7-10
µm. Solid black curves show corresponding black body temperatures.
DATA COLLECTION
We collect:
•Zenith sky temperature measurements using in-
frared thermometers [2].
•Precipitable water data from radiosondes at Al-
buquerque (ABQ) and El Paso (EPZ), launched by
the National Weather Service.
Figure 2: Time series of measured temperature (red) and precipitable water
(blue).
CURRENT ANALYTICAL RESULTS
Figure 3: (Top Left): Temperature and TPW based on averaged ABQ and EPZ ra-
diosondes. (Top Right): Temperature and TPW based on averaged 12Z and 00Z
radiosondes. (Bottom): Temperature and TPW based on total mean. The blue,
orange, and green circles correspond to results of Figure 2. Figure 4: Pac-Man residual plot.
CONTACT INFORMATION
Spencer Riley: spencer.riley@student.nmt.edu
Vicki Kelsey: vicki.kelsey@student.nmt.edu
Documentation Page: https://git.io/fj5Xr
ONGOING RESEARCH
We plan on using machine learning techniques to pre-
dict whether or not a day is overcast or clear sky.
Figure 5: Current machine learning network configuration.
We are currently developing a method to automate the
temperature measurement process. This wold allow
more consistency in the data collection process in ad-
dition to the potential of deployment anywhere
ACKNOWLEDGEMENT
We would like to thank the Institute of Complex Addi-
tive Systems Analysis, and the New Mexico Tech Re-
search, Physics, and Student Affairs departments for
funding our trip. As well as Dr. Kenneth Minschwan-
ner for his invaluable support of the project.
DISCUSSION
Using the computational tool, we have experimentally
verified the exponential relationship between precip-
itable water and zenith sky temperature.
We also have determined that there is a moderately
strong correlation between the amount of precipitable
water and the temperature at the zenith (R2= 0.698).
The results of our analysis show that the data fits
within ∼1σ, as seen in Figure 4.
We suspect that the sources of error include:
•Slight variations of measurement time and loca-
tion
–Temperature measurements taken at about
1100 in Socorro while measurements in Al-
buquerque and El Paso are taken at 0500 and
1700 local time (MST)
•Atmospheric phenomena may have introduced
bias into the temperature readings
COMPUTATIONAL METHODS
We have developed an open source tool in R for ana-
lyzing the relationship between precipitable water and
temperature. The tool utilizes the numerical methods
•Linearization of an exponential relationship
•Linear regression analysis
The computational tool contains four plot sets to visu-
alize the data collected:
•Time series of average temperature measure-
ments and precipitable water readings
•Analytical plots that show the correlation be-
tween temperature and precipitable water
•Individual sensor plots that shows the time series
of temperature measurements for each of the in-
frared thermometers.
•Charts that show the distribution of observation
conditions recorded by each of the infrared ther-
mometers.
