Science Buddies: "Ask an Expert"

Dear sirs,
I need information on how to obtain an inexpensive GPS that I can turn on and off the WAAS and has capability for longitude, latitude, and perhaps altitude. Any suggestion will be appreciated. Thank you.

This might do it
http://www.amazon.com/Bad-Elf-Performan ... d+Elf+1000

Dear Sir

Thank you for your help in picking out a GPS. I picked out a Garmin Nuvi 350. I have collected six weeks worth of data and today after several calls to Garmin, revived additioaal information. I do not believe I can use my six weeks worth of data. Should I keep this data in my log book and continue on a fresh page or begin again in a whole new book? I now have access to better data collection. I feel depressed. Thank you for any feedback.

bobbyo,

I know it can be frustrating to do so much work, only to find an error that renders it unusable. However, I would advise never throwing data out. You don't mention why you can't use it, but it does constitute work done for the project and even mistakes themselves provide information. I would just write a paragraph describing what happened and why the previous work isn't usable for the project and then start on a new page, probably with a paragraph stating how you fixed whatever caused an error in data/data collection.

Dear Sirs,
My project has ran into another issue. I chose this project because the sun was supposed to be in the greatest 11 year cycle of geomagnetic storms this past fall but as it turns out, it was a very very quiet cycle. My experiment involved working backwards from my GPS and recording the data of the long, lat, and elevation readings with WAAS on and WAAS off. My hypothesis was based on different reading of the WAAS on vs WAAS off for the 3 data groups I was collecting. I have collected data 2x a day from Nov, Dec and up thru today. There has been a rare difference in readings and more unfortunate, there has been no geomagnetic storms Kp-5 or greater since I started my data collection. I will continue collecting data but what should I do if there is no Kp-5 or larger geomagntic storms on the sun? Scrap everything?, do something else? Will an experiment based on something that is very uncontrollable be thrown out of a science fair? My project was based on a similar project. The experiment made the hypothesis, that by knowing there was a Kp-5 or greater geomagnetic storm, your GPS readings would be altered. Mine is the oppposite, working backwards. Altered GPS readings could show Kp-5 or greater geomagnetic storms. Please let me know if this experiment is a no go. I have collected some altered GPS readings, but very rarely, and none correlated with a geomagnetic storm, but there happened to be radio backouts on these rare occassions. I would note this but still NO GEOMAGNETIC STORMS!!!! Bobbyo

I wanted you also to know my original collection data issue. For a couple of weeks I took GPS readings from my Nuvi I got on EBay. I called Garmin and received instructions for the older Nuvi and was told where I could get my longitude and Latitude readings. The longitude gave readings such as N 39 29.894
Latitude W 076 27.231
Elevation 48 ft.
I recorded data for 2 1/2 weeks.

I asked what type of readings these were and how do I convert them to readings such as N 39 29 53.2. W 074 27 13.0

Garmin did not know why the readings were in this format. After an hour on the phone, the tech showed a different area in the Nuvi that gave me the longitude and lat readings in the format I asked about. Are the 2 1/2 weeks worth of data collected the first way still able to be used? Are the first longitude and latitude readouts just displayed differently. Thanks
Appreciate your help very much.

Hello,
Figured out my issues. Did a lot of research and asked a lot of questions to Garmin and the NOAA. I am seeing something but still need to graft. I will be making three grafts one for each magnitude of error vs Kp values for the geomagnetic storms. I also noticed some odd things dealing with solar radiation storms and radio blackouts when errors occurred on my GPS. Would you advise to note if there were high solar radiation storms and radio blackouts or just concentrate on Kp values. I thought I would note that these occurrences that could also be studied for GPS errors. I believe I have data to do all three but my hypothesis is based solely on geomagnetic storms.
Thanks Bob

Dear Sirs,
I am having trouble graphing my data. I have collected data since October. I have collected 2x daily Kp values and error calculations for longitude and latitude readings. Kp values can range from 0-9 , mine range from 0-5. Error calculations range from 0-1.0.
I want to use scatter charts for latitude and longitude. I made scatter graphs but they don't look good. I think it's because there is a lot of redundancy. Most data points are Kp values of 0 and error calculations of dirty 0, 0.1 or 0.2.
Should I use another type of graph? I want correlation studies done and thought scatter was the best between 2 variables.

In my experiment, I also made a guess as to whether there was a geomagnetic storm based on error calculations. I then looked up Kp values and determined if I was right or wrong. I was thing of doing simple percentages of correct and incorrect guesses per all the data. Should I make a graph for the guesses? If so, which graph would be the best to use?

Please help Thanks

Hello. This sounds like a very interesting project. I am not a geomagnetic storm or GPS expert, but I will try to help as much as I can.

To answer your Lat / Lon coordinate question. Lat / Lon can be described in different formats. Different programs and navigation systems use one of these formats:
Degrees, minutes, seconds (N 39 29 53.2)
Degrees, minutes, decimal minutes (N 39 29.8867)
Degrees, decimal degrees (N 39.4981)

There are 60 seconds in a minute and 60 minutes in a degree. Start from right to left to convert deg, mins, secs to degs, mins, decimal mins and then to degs, decimal degs. Let’s use the number I listed above.

Starting with degrees, minutes, seconds: N 39 29 53.2 is 39 degrees north, 29 minutes, 53.2 seconds. Starting from the right, we want to convert 53.2 seconds to minutes:
- 53.2 seconds / 60 seconds * 1 minute = 0.8867 minutes
- Add 0.8867 to your 29 minutes => 29.8867 minutes
- You can now state the same Lat coordinate as N 39 29.8667 (degrees, minutes, decimal minutes format)

- Move on in your conversion. We want to convert 29.8667 minutes to degrees:
- 29.8867 minutes / 60 minutes * 1 degree = 0.4981 degrees.
- add 0.4981 to your 39 degrees.
- You can now state the same Lat coordinate as N 39.4981 (degrees, decimal degrees format)

Therefore, N 39 29 53.2 is equivalent to N 39 29.8867 is equivalent to N 39.4981!
The N simply indicates this location is north of the equator. There is an equivalent location on the earth south of the equator. This would be listed as S 39 29 53.2.
More information on lat / lon coordinate system and formats here:
http://www.maptools.com/tutorials/lat_lon



I am sorry your data results aren't what you were hoping for. This is not uncommon in science experiments. That doesn't mean you did anything wrong, it doesn't mean you didn't learn anything, and it doesn't mean there isn't useful information. I agree with theborg…you should never throw data out or ignore it. It is a record of work you did and problems that have already been investigated. By reporting this data, this could save another scientist (yourself included, perhaps!) time / money / effort in the future by not having to repeat what you have already done.

You say Kp values should range 0-9 but yours range 0-5. Again, I am not familiar with this rating system, but the NOAA website describes values ranging from 5-9. They list a G scale from 1-5. Are these the numbers you are referring to?
http://www.swpc.noaa.gov/NOAAscales/index.html#

Scatter plots are an excellent way to investigate correlation. Why are you interested in plotting lat / lon? Plotting lat / lon coordinates will simply show where you took your readings. I assume you took all of your data readings from the same location? Your back yard or at school perhaps? If you took your readings in the same place (or pretty close…as in within the same city), I’m guessing your chart consists of several points on top of each other.

How about a scatter plot consisting of your kP value on one axis and your error calculations on the other axis? Have you been following this procedure? In the procedures section, they describe a recommended scatter plot that I think will work well with your data.
https://www.sciencebuddies.org/science- ... p009.shtml

So, you don’t think your data is interesting, because there haven’t been severe geomagnetic storms during your data collection period. There is still valuable information in your results.
- What was the range of storm severity during your data collection period? You can report to navigators that as long as geomagnetic storms are less than XXX severity (the highest severity of storm that occurred during your data collection), then the error reading between WAAS and no WAAS shouldn't be greater than XXX (your greatest error calculation). Be sure to report your sample size. It sounds like you have a lot of data. The more data…the higher confidence in your results.
- You mentioned solar radiation storms and radio blackouts were perhaps related to errors in your GPS. Another plot you could do is error reading on the y-axis over time on the x-axis. Since you say your data is pretty benign, I’m thinking your data plot will be close to a straight line across your data collection period. Are there any data points that spike up higher or lower than the rest of the data points? If so, on what day / time did these spikes occur? Do these dates / times correlate to a reported solar radiation storm, radio blackout, or other space weather phenomenon? If so, then that will be important and interesting data to report!
- I like your last idea about trying to predict that a storm happened, based on your error reading. Plot it out, calculate the percentages as you suggest, and see what you get! Bar graphs are a nice way to report percentages.

Please don’t be discouraged by the fact you don’t have as exciting data as you were hoping for. Your experiment and results are still important to the scientific community. Report as much information as you can think of. Although it may not sound interesting to you at this moment, it could be a critical piece of information to another scientist reading your results.

Please write back if you have further questions. Keep your chin up, smile, and be proud of your experiment!

Thank you for your response.

A Kp of 5 is the lowest level storm to make the NOAA scales category. Although four is also mentioned due to storms possible interferences on other parts of life, such as homing pigeons. Geomagnetic storms go to G1-G5 and are matched with the Kp values of 5-9. Zero is understood as the Kp value that is below 5 or 4. All of my results had a Kp of 0, 4, 5, 6, or 7. The vast majority of the Kp values were 0. The GPS error calculations ranged from 0 to 3.3. The 3.3 was an outlier. After doing a correlation study, there seemed to be no linear correlation.

Do you believe this data should be shown on a bar or line graph? I did a scatter chart and it doesn't look very good because of all the redundant data. This experiment could not prove that a person could calculate GPS errors and say there was a geomagnetic storm occurring of a Kp 5 or greater. I plan to note other space disturbances that may have interfered with my error calculations. I plan to make these interferences into charts and percentages.

Is there anything else that you can reccomend?
Thank you
Bobbyo

I still recommend a line graph or scatter chart. Bar graphs are useful for summing up data, but I think it's better if you plotted kP value v. GPS error. You can't do that in a line plot. If most of your kP values are zero, was there any interesting phenomenon on the days where the kP values were 4, 5, 6, or 7? Since you said there was no correlation, I'm assuming the answer is no.

- Have you looked at historical data? Can you find out how much GPS error there is for days with kP values of 4-7 (your range of non-zero data)? Do those results match your data results?
- On average, how many geomagnetic storms occur per year? Is there any kind of annual cycle where geomagnetic storms are more prevalent? If these storms are more popular during the spring / summer / fall time frame, this could explain the more quiet times during your data collection. Solar cycles occur about every 11 years. Is this related to geomagnetic storm activity?

If there are still no interesting results, all you can do is report your findings and say there is no significant data. This happens some times in experiments. You collected a very small amount of data. Collecting data since October sounds like a lot of data, but in terms of studying weather effects, that is actually a very small amount of data. That is not your fault. You could very easily have collected data during a very active time. This winter just happened to be a quiet time...you can't predict that.

Your conclusion discussion should include:
- No statistical significance / correlation was found between GPS error readings and geomagnetic storm activity. You didn't prove or disprove anything. You just didn't find statistical significance. This may sound the same but it is not. You are taking a small amount of data hoping it will represent the greater population of all geomagnetic storms over time. In experiments, there are statistical calculations used in test planning to maximize the probability that your experiment results reflect the true results for the greater population. However, that is well beyond high school math, and that was not required for this experiment.
- Display as much data as you can, even though you may not find it interesting or significant. Showing data backs up your statement that there is no correlation.
- Do you think there is anything you could have done different in your experiment that may have lead to different conclusions?
- If you were to continue with the experiment, would you do it the same way, or what would you change? Would you keep the same hypothesis, or draft a new one? Scientists sometimes use "failed" experiments as a stepping stone to generate new hypotheses.

Be sure to read the section here on conclusions. It gives great suggestions of how to present data that didn't support your hypothesis.
https://www.sciencebuddies.org/science- ... ml#keyinfo

Thank you for all your help. I did a space weather event pie chart for all activities that coincided with GPS error calculations. 41% of my elevated error calculations had no known reason. Percentages were given for Kp, sudden geomagnetic impulses, electron and proton flux, type 2 and 4 radio emissions, and x-ray flux r2 and r3.

Kp values of five and above occurred only 10 %of my readings.
14 % occurred when type 2 radio emissions were happening. The rest of the percentages were lower then 14 %

I did find interesting data.

I chose the fall of 2013 because it was supposed to be the most active in the 11 year cycle...they lied

I took your advice and looked up statistics on GPS errors and Kp values. I found that DST values are more reliable then Kp values. This would be another science fair project.

I did a pie chart for percentages and did line graph for each Kp value for longitude and latitude. I now have a better understanding on my project.

Do you recommend I present the correlation formula or just tell them I used the Excel's programming. I never had a statistics course but I understand the positive, negative, or 0 correlations.

Thank you
Bobbyo

Hi again,
Wanted to verify my variables. I could control the place and times and whether WAAS was on or off when I took readings. These are controls. I used my error calculations as my independent variables because I chose the time and place ,WAAS on/off for them. I made the Kp values my dependent variables because after I made my guesses to whether there was a geomagnetic storm was occurring, I looked up the Kp value and determined if I my guess was correct. I was trying to prove errors were caused by Increased Kp values. I'm confused. Should I reverse my variables. I am working backwards. Other experiments learned of Kp values first then made error calculation and They had Kp as independent and error as dependent. I'm doing the opposite.

Out of 92 data collections , only 9 occurred when geomagnetic storms were occurring.
7 Kp 5 , the lowest magnitude storms were predicted correctly only 14% of the time. 2 Kp 7 storms were predicted 100% . But there were only 2 Kp 7's. I don't think I can use only two data results to make a true statement. This was a very quiet time for geomagnetic storms and I only captured the 9 Kp values that were 5 or greater.
My hypothesis stated that I could predict geomagnetic storms but At 14% correct based on 7 vales and the 100% based on 2 values, I don't think my experiment proved my hypothesis correct.
As far as my CONTROLS , place and time of readings
As far as VARIABLES: INDEPENDENT. GPS readings and Kp values.
DEPENDENT. Predictions made
Please give your feed back.
Do I look like I'm on the right track?

Hi. I apologize for my delayed response.

Remember, an experiment whose conclusion does not match your hypothesis is NOT a failed experiment. This happens quite often. A successful experiment is one that presents useful data. However, in cases where the results don't match the hypothesis, it is even more important to discuss in your results / findings / conclusions what you researched, why you set up the experiment the way you did, what you would do differently if you could do the experiment again, and what experiments could be conducted in the future, based on your test results. You have mentioned several good points in your recent posts:
1. I like your space weather events pie chart. It demonstrates you examined different space events (not just geomagnetic storms) that could have affected your test results
2. You chose your data collection period based on predicted peak performance of the 11 year sun cycle. This demonstrates you using your research in attempt to set up your experiment for the highest probability of success. Even then, as you now see, mother nature can be unpredictable!
3. DST values versus kP values. Sounds like a great future project!

As far as presenting the correlation formula v. the Excel programming. Why not discuss both? Did they present the same result? At the high school level, you won't be expected to have performed complicated statistical formulas. But, if you used the correlation formula, this demonstrates you attempting to use advanced methods, which may win you points at the science fair.

Your controls sound fine. I agree that you should switch your independent and dependent variables. Dependent variables are those that are affected by changing levels of your independent variables. You are studying how GPS errors are affected by kP values (intensity of geomagnetic storms).

You are correct that you shouldn't make conclusions on based on 7 or 2 values. In more advanced statistics, you can actually calculate confidence behind these findings, based on number of data points; but those calculations are well beyond high school math. Present as much data as you can in the different ways you investigated the data (your pie charts, line graphs, etc), and state exactly as you said...the current data does not support your hypothesis.

You never prove or disprove a hypothesis, even if you had years worth of data. Remember, an experiment typically only uses a small sample of data you hope correctly represents what would happen if you could have the entire population set of data. In your case, that would mean collecting data all the way back to the beginning of GPS devices. Since an experiment typically only contains a portion of the population data, a scientist always runs the risk of the test results not matching the true conclusion of the entire population. These are known as Type I and II errors, better known as false positives and false negatives (advanced statistical concepts you don't need to worry about for your experiment). When employing more advance statistical methods in an experiment, the scientist uses statistical calculations in attempt to minimize the probability of the experiment results not matching the true population outcome. Because you always run this risk, it is not correct to say you proved or disproved your hypothesis. Instead, you say you either rejected your hypothesis or failed to reject your hypothesis. It may sound the same as prove or disprove, but in the world of statistics, these statements are different.

I would like to thank everyone who helped with any questions I had for my project. I received an honorable mention. I again appreciate all of your help.

Congratulations!!! We are very proud of you. I know it's hard when your data doesn't work out as planned, but you now see that unplanned results are still useful results. Well done.

Hi Bobbyo,

I am planning to do a similar project but I am not finding a GPS model in which I can turn off/on the WAAS feature. Can you please tell me the GPS model you used for this project?

Thanks,
Bharat