X - Sites at Penn State

advertisement
Statement of the Problem: Suppose your boss gives you the day off, and a beach day is long
overdue. Will Mother Nature cooperate with your plans, or will she dampen your beach day
festivities? Sure, you could check the latest weather report, but there are so many variables to
compile in such a short time---not to mention the many others ignored. It would be nice if there
was a simple program that took all of your worries into consideration and gave you a straight up
answer to the age-old question: Is it a good beach day?
Objectives: We want to give the public information on current and future weather conditions that
will affect the quality of a beach day. The factors are to be combined together in order to give an
overall view of the quality of the day. Based on the information given, the public can decide
whether the conditions are favorable to go to the beach. The factors under consideration are wind
speed, sea surface temperature, heat index, cloud cover, probability for precipitation and
probability for convection. Safety information will also be given based on air quality and potential
for severe weather. We will output a rating that the user can interpret and reach the conclusion
that it is or is not a good beach day.
Description of Input Data: The data used in presenting this information includes SREF files from
ensemble members. There are 21 ensembles with each ensemble having 30 time stamps. These files
have forecasted conditions across the US. There are data on temperature, relative humidity, and
components of the wind at different height levels, as well as CAPE, precipitable water, and mean sea
level height. There are data on particulate matter (2.5 μg) and ozone taken from AirNow. Data on sea
surface temperature are taken from NOAA. There are some limitations to these datasets. The SREF files
are ensemble forecasts; they do not guarantee future weather conditions. These data are only updated
every three hours, which will not allow for accurate data when significant mesoscale events are
occurring. Another limitation is sea surface temperatures are taken on a weekly basis. To activate these
data archives, the user will be prompted to input the following information: date, time, and location.
Locations are limited to the 20 most popular beaches in the US.
Decision Support Algorithm: Archived data corresponding to user information is gathered from 21
different ensemble members. Data are then run through a series of appropriate functions (see Figure
1). The variables returned from these functions are combined as a weighted average, which, in turn, is
returned to the user for interpretation. A more detailed explanation of this algorithm follows.
Implementation Details: Once the code is run, the user will be prompted for a year, month, and day.
They will then be asked to select a coast and choose from a list of beaches. At this point, the code will
start making calculations through functions. There are multiple functions defined for this code. There is
a function to calculate each individual factor of the beach day decision (see Figure 2). There are seven
categories (wind speed, air quality, sea surface temperature, heat index, cloud cover, precipitation, and
storm potential) to weigh in the determining the quality of the day. Each of these seven categories has a
probability criterion within their functions. This probability (0-10, with 0 being bad and 10 being good) is
a score that decides how different values of the factor will affect the quality of the day (Figure 3). For
example, if there is no precipitation in the forecast, the day is assigned a score of 10. If there is a 90 %
chance for precipitation, the day is assigned a score of 0. Each of the seven functions run through a loop
of all 21 ensemble members for the specified time and location (input from user). The calculations in
each function are done for each of the 21 ensemble members and they are then averaged to give one
value. That value is then assigned a score (see Figure 3). At the end of this loop, there are seven values
(one from each function), as well as the score assigned to that value. An equation takes a weighted
average of all these scores (see Figure 4). Because of how CAPE varies around tropical storms and
hurricanes, there are two different averaging equations. CAPE (storm potential) is included only if there
is less than a 90% chance of precipitation (Figure 4a), otherwise, it is excluded (Figure 4b). All of the
factors are weighted based on importance of that factor. For example, chance for precipitation is
weighted heavier than wind speed because you are less likely to go the beach when it is raining than if it
were just windy.
Description of Output: The output is a series of print statements and a plot. The initial statement is the
calculated weighted average of the scores. This is followed by statements for temperature and heat
index, ocean temperature, cloud cover, precipitation, and storms. Because this program loops through
just one time period, these statements are only valid for the 3 hours following the user inputted start
time. The plot shows a visual comparison of the individual factors affecting the quality of the day, as
well as the weighted average. This is plotted as a bar graph, with a 0 to 10 scale. Again, 0 is a poor day,
and 10 is an excellent day.
Test Plan and Examples: The test plan is a day with conditions that would be favorable for going to the
beach. Figure 5 shows how that day would be rated. A second test plan, Figure 6, shows how a poor
weather day would be rated. Figure 7a is a sample output for August 2, 2011 at 15:00 for Sea Island
Beach, GA. Figure 7b is the plot that corresponds to Figure 7a. Figure 8a is sample output for August 2,
2011 at 15:00 for Miami Beach, FL. This location has potential for severe weather at this time, and
therefore has an additional output statement. Figure 8b is the plot that corresponds to Figure 8a.
Potential Uses of this Product: This product is for beachgoers, vacations, and beach-based businesses.
Besides the obvious use for deciding whether the day will be a good beach day, business owners can use
the product to associate a good beach day with more potential business. It can also double as a safetyrisk index for beach locals.
Future Improvements: Improvements can be made in streamlining the functions. All of the function
files could be combined into one file instead of having 4 separate files. There are also many elements
that could be added to the program. Data concerning wave height, population, UV index, and rip
currents could be included. Additional beaches could also be added to the code so the code can be used
by more people. The precipitation score could be made more proportional to the amount of
precipitation that will fall as opposed to just the whether or not it will rain.
Figure 1 Functions data must go through.
Function
Wind Speed
Air Quality
Sea Surface
Temperature
Heat Index
Cloud Cover
Precipitation
Storm
Potential
Variables
u and v
components of
wind (m/s)
Air Quality
Index values for
PM2.5 and 8hr Ozone
temperature of
sea surface (◦C)
surface
temperature
(K) and surface
relative
humidity (%)
relative
humidity (%) at
850 mb
total
precipitaton
(kg/m2)
CAPE (J/kg)
Calculations
Wind speed is calculated using the wind
speed equation: sqrt( vu2 +vv2)
Probability Decision
Based on comfort
No calculation
Based on health
standards set by the
NAAQS
Convert to Fahrenheit
Based on comfort
The heat index equation can be found at
www.hpc.ncep.noaa.gov
Temperature is converted to Fahrenheit.
Adjustments equations are used if the
relative humidity and temperature fall
above or below certain thresholds. These
adjustments can be found at the same
website.
Percentages of relative humidity at 850mb
were related to what the likely percentage
of cloud cover would be at that height.
Chance of precipitation calculated by
averaging the probability that there will
be precipitation across all ensembles.
No calculation
Based on thresholds
stated at
www.srh.noaa.gov
Based on comfort
Based on comfort
Threshold values based
off of information found
at
www.balloonacyltd.com
Figure 2 Table of functions used within the code.
Variables
Wind Speed
Air Quality
Sea Surface
Temperature
a) Atlantic
Probabilities within the Function
0-5 = 10/10
5-10 = 7/10
10-15 = 3.5/10
15+ = 0/10 [units: m/s]
0-50 = 10/10
50-100 = 8/10
100-150 =6/10
150-200 = 4/10
200-300 = 2/10
300+ = 0/10
a)
85-80 = 10/10
85-90 or 75-80 = 8/10
70-75 = 6/10
b)
65+ = 10/10
60-65 = 9/10
Ocean
b) Pacific
Ocean
Heat Index
Cloud Cover
Precipitation
Storm Potential
90F-95 or 65-70F = 4/10
60-65 = 3/10
55-60 = 2/10
95-100 or 50-55 = 1/10
100+ or <50 = 0/10 [units: ◦F]
< 90 = 10/10
90-100 = 7/10
100-105 = 3/10
105+ = 0/10
[units: ◦F]
0-50% = 10/10
50-75% = 7/10
75-100% =5/10
0% = 10/10
0-25% = 9/10
25-50% = 8/10
50-75% = 7/10
75-90% = 4/10
90-100% = 0/10
0-300 = 10/10
300-1000 = 7/10
1000-2500 = 5/10
2500-3500 = 2/10
3500+ = 0/10 [units: J/kg]
55-60 = 7/10
50-55 = 5/10
< 50 = 0/10 [units: ◦F]
Figure 3 Score assigned to each variable out of 10. 0/10 means it is not a good day to go to the beach based solely on that
variable, and 10/10 is a great day to go to the beach based solely on that variable.
Figure 4a. Weighted average score
including storm potential
Figure 4b. Weighted average not including
storm potential
Figure 5 Example of how a nice summer day would be rated
Figure 6 Example of how a poor weather day would be rated
Figure 7a Sample output for Sea Island Beach, GA on 8/2/11 at 15:00
Figure 7b Sample plot for Sea Island Beach, GA on 8/2/11 at 15:00
Figure 8a Sample output for Miami Beach, FL on 8/2/11 at 15:00. Note the additional statement concerning potential for
severe weather.
Figure 8b Sample plot for Miami Beach, FL on 8/2/11 at 15:00.
Download