Science Buddies: "Ask an Expert"

Thanks again for the help. Today I started recording data every three hours. The times I have chosen are the universal times converted into the time for my time zone. So far, it's helped me a lot with knowing when to expect space weather activity. I noticed that when I recorded IGP delay in the middle of the day today, the number was the highest in the data i have so far, even though there was no space weather, the proton flux was flat, and kp was below 4. I was wondering if you could explain to me what factors can affect how IGP delay varies other than geomagnetic storms and the other factors you mentioned in earlier posts.
Thanks,
- coolgirl285

I have one more question: Do you know if solar activity for the rest of this month and January, (during the course of my experiment), is high? I don't want to do my experiment if the sun's activity isn't high enough.
Thanks,
- coolgirl285

Hi coolgirl285,

Your solution to the time zone difference sounds great! I'm glad to hear that the project is starting to become clearer for you.

The IGP delay is a function of activity in the ionosphere. Solar activity definitely affects the ionosphere, so the same processes that give rise to solar radiation storms, radio blackouts, and geomagnetic storms also affect the ionosphere. Lightning can also cause ionospheric disturbances. The "Ephemeral ionospheric perturbations" section of this Wikipedia article describes these factors in more detail:

http://en.wikipedia.org/wiki/Ionosphere

This article from Stanford University's Solar Center is also quite helpful for understanding the ionosphere in general: http://solar-center.stanford.edu/SID/ac ... l#overview

If there were a lightning storm in the place you were recording IGP delays for, it's possible the lightning could disturb your measurements. Building up observations through time will help you see through the noise and complications to discern the relationship between Kp and the IGP grid delay. (If lightning is contributing, it's very unlikely that you will be recording data during a lightning storm every day!). One thing you could try, if you want, would be to average three IGP measurements each time you record data. IGP delays are updated every 3 minutes, so you could write down the IGP delay at (for example), 6:00, 6:03, and 6:06am, average those three delays, and use that average value in your analysis. But, I don't think that is really necessary. However, you can try it if you want to.

Solar activity see-saws up and down, but right now we are in the solar maximum. This is the most active the Sun will be for another ~11 yrs. You picked the best possible time to do this project! The sites below provide some more details on the current solar cycle and predicted solar activities.

http://solarscience.msfc.nasa.gov/predict.shtml
http://en.wikipedia.org/wiki/Solar_cycle_24

Let me know what other questions you have, and how else I can help you.

Keep up the good work! Your dedication to and perseverance through this challenging project is remarkable.

I looked at predicted kp values from this link: http://www.swpc.noaa.gov/products/27-da ... ic-indices and it looks like there will only be one geomagnetic storm from December 22 to January 3. I'm entering my city's regional fair as an independent, so I can do my experiment a little longer if I need to, since I don't really have a deadline. The longest I can do my experiment until is around February 21 (which would be 2 months after I started my experiment), because even though I don't have any deadlines, I'm supposed to be done before March. What should I do if there aren't enough geomagnetic storms for me to compare to igp delay?

Hi coolgirl285,

That's a really good question! A few things come to mind.

1) There is a lot of time between January 3rd and February 21st, so even if only one storm happens between now and the 3rd, more storms will probably happen before February 21st. Even if there were only one storm every two weeks, you would still have data on 4 geomagnetic storms by February 21st.

2) You can assess the relationship between the IGP delay and planetary K index even when a geomagnetic storm isn't happening. Think about "regular" weather: we talk about the weather regardless of whether a hurricane or heat wave is happening. Even during a nice week with no rain storms, the weather will be slightly different each day (a bit warmer or cooler, slightly different humidity, etc.). Likewise, space weather is happening even when a geomagnetic storm isn't going on. You will still see changes in Kp and the IGP delay over time. For example, let's say that Kp varied between 1 and 4 over the period of a week. That's not enough to qualify as a geomagnetic storm, but Kp definitely varied during that time, which you could relate to variations in the IGP delay.

3) It's important to establish the "baseline" for your study. Whenever a research project looks at how particular data sets respond to outside events, a crucial part of the study is determining what the data look like when an event isn't happening. Scientists call this the baseline or background. Without understanding the baseline, it's impossible to know what an event (a geomagnetic storm, in your case) looks like compared to normal. So, you actually need to collect data at times when a geomagnetic storm isn't going on so that you can compare storm and non-storm conditions.

4) Let's take a worst case (and highly unlikely) scenario—there aren’t any geomagnetic storms, radio blackouts, or solar radiation storms between today and February 21st. Does this mean your project is a failure and won't do well at a science fair? The answer is absolutely no--your project is still a success and can do well at a science fair. Why? Because you followed the scientific method: you asked a question, came up with a testable hypothesis, collected data to test your hypothesis, analyzed your data, and came to your own conclusions based on your data. (You would be surprised by the number of projects that draw conclusions that aren't based on the data collected by the student.) Your conclusion might be "I can't conclusively tell whether my hypothesis is true or false because no geomagnetic storms happened". That's a perfectly valid conclusion, and it's based on your data. Judges and teachers understand that Mother Nature doesn't always cooperate. You will still be able to present data on the relationship between Kp and the IGP delay when geomagnetic storms aren't happening, which is interesting in and of itself. What matters is how carefully you followed the scientific method and how detailed your data collection process was. That being said, it's exceedingly unlikely that you won't get at least a couple of space weather events between now and February 21st.

Post back with more questions! It sounds like this project is going pretty well for you right now. And, you're asking really good questions that show you are thinking deeply about your project. That will take you a long way when it comes to writing about and talking about your project.

I don't quite understand what you said about baseline. Since only one geomagnetic storm has showed up in my data (one that occurred today), the rest of my data was recorded when space weather events weren't occurring. So, would the baseline for my experiment be how the kp varies when geomagnetic storms aren't occurring?
Thanks,
- coolgirl285

For my experiment, I have been keeping track of any changes in the planetary k index and GOES proton flux, so that I know when to expect a storm to occur. I know that the planetary k index is for geomagnetic storm activity and the GOES proton flux is for solar radiation storms, so, I thought I should also note any changed in the GOES x-ray flux for radio blackouts: http://www.swpc.noaa.gov/products/goes-x-ray-flux but I don't know how to interpret it and which of the two graphs I should be refer to for my observations.
Thanks,
- coolgirl285

Hi coolgirl285,

Oh dear! I didn't get an email from the system when you posted last! I'll double check my account settings to see if the notifications got switched off or something. I'm so sorry it's been so long since you posted!

Correct, the baseline for your experiment is how much Kp varies when geomagnetic storms aren't happening. I drew a couple of graphs to hopefully clarify the concept. You can see them by downloading the "baseline.pdf" file from https://sites.google.com/site/scibudkpe ... mple-graph. (My attachment quota on the Forums is full, so I'm linking to an external public website.) This data is totally fictionally--I made it up to make a point. The x-axis of these graphs is time; the y-axis is Kp (again, these are totally made up numbers). I put in a pretend geomagnetic storm from August 11-14 (green box).

The graph of the left is an example of a dataset with a noisy baseline. There's a lot of variation in Kp even when a storm isn't happening. This makes it very difficult to tell when a storm is happening--Kp does stick up high enough above the background chatter in the data to be very obvious. The graph on the right is an example of a quiet baseline. Kp doesn't change very much when a storm isn't happening, so it's very easy to see the storm in the Kp data--it shows up as a big hump. Collecting data when storms aren't happening helps you know how much chatter is in the Kp data when a storm isn't going on.

Checking the GOES x-ray flux is a good idea. Use the graph on the right side of the page when you collect data--that graph shows how the x-ray flux (# of x-rays of a particular energy that pass through a square meter every second) for the past 6 hours. Focus on the red line. Solar flares will show up as big spikes in the flux. Any data that spike up above the 10^-5 line on the y-axis indicate an event large enough for NOAA to issue a radio blackout warning (see the "radiation blackouts" section of http://www.swpc.noaa.gov/noaa-scales-explanation).

Do I already have baseline data for my experiment since I have plenty of data recorded at times space weather isn't occurring? In the procedure for the experiment I was originally following when I was using gps errors: https://www.sciencebuddies.org/science- ... p009.shtml it says to spend 3 weeks collecting baseline data before actually starting the experiment. I'm confused by this; should I have collected baseline data for kp before starting my experiment?
My experiment has run into another time issue. I started my experiment during winter break, so recording data wasn't an issue. School starts this week and three of the times i need to record data at are during class. I gave my mom the instructions for recording data so she can help me out for a few days. I don't think my mom can record data at those three times for my whole experiment though, since she runs arraigns while I'm at school. I'm going to ask my teachers if I can use my phone to record data, but this also wouldn't be something I can do everyday (during tests/quizzes for instance). I know this isn't necessarily a science question, but any suggestions would be extremely helpful.
Thanks for the help,
- coolgirl285

Hi coolgirl285,

Great questions. Yes, you already have baseline data for your experiment. Every time you make an observation and a space weather event isn't happening, you are recording data for your baseline. You do not need to collect baseline data separately before starting your experiment.

As far as recording data goes, there's not a 100% "right" answer. Having your mom help you is definitely a good idea.Asking your teachers if you can use your phone to record data is also a good idea. When you talk to them, be sure to explain that you are doing a science project and need to collect data (i.e., you're doing science, not checking out Instagram). You might talk to your science teacher first and see if he or she talk to your other teachers to get permission for you to use your phone. Your other teachers might respond better if your science teacher, instead of you, asked them. Another thing to keep in mind is that if you miss an observation occasionally, it isn't the end of the world. You definitely want to stick to your data collection schedule as much as possible, but if you have to make some adjustments (e.g., record data at, for example, 12:15 instead of 12:00), that's OK. Just be sure to jot down a note about it in your lab notebook when you make an adjustment. For example, you could adjust the times so that you are checking for data during your lunch break or while passing from one class to another.

Keep up the good work!

During the past few weeks, I haven't collected as much data as I had hoped. I missed various geomagnetic storms and radio blackouts. Recording data in school was my main issue. I talked to my teachers, but they said that they can't let me leave the class so many times in one day. My mom didn't understand my experiment at all, but even if she did, she leaves the house while I'm at school to pick up my sister from daycare, go grocery shopping, etc. I'm really worried because my data from the past 2-3 weeks has a bunch of missing times. I thought I could do my experiment longer, since I'm entering my city's fair as an independent, but it looks like need to be done my project (not the display) in the next week or so.

I'm also having issues with what to do with my data. Aside from calculating the correlation coefficient like it says to do in the procedure, how should I analyze/display my data? Sorry if it seems like I'm trying to get someone else to do the work for me. I usually ask my science and math teachers for help with this part.
Thanks for the help,
Coolgirl285

Hi coolgirl285,

I'm sorry it has taken me so long to get back to you. I've been traveling the past couple of weeks. I had hoped someone else would be checking and responding while I was away, but apparently that wasn't the case. I hope this information isn't too late, particularly since your project is due so soon.

As far as data goes, use the data you have and don't stress about the rest. You did the best that you could - putting a lot of effort into getting the data - and that you show. If you get asked why there are gaps in your data, simply explain the reasons. Judges/teachers should understand.

As far as data analysis goes, I suggest making a few different graphs. First, make a graph with time on the x-axis and Kp on the y-axis. This graph will show how Kp varied over the time you collected data. Second, make a graph with time on the x-axis and IGP delay on the y-axis. This graph will show how the IGP delay changed over time. You may find it helpful to display these graphs near each other on your display board so that people can easily see how they relate. Third, make graph with Kp on the x-axis and IGP delay on the y-axis. This will clearly illustrate the relationship between these two variables. When you put the graph on your display board, you can write/type the correlation coefficient on the corner of this graph.

When you make the third graph, look carefully to see if there is a change in the IGP delay at Kp indices >5, which is the lowest Kp index that still counts as a geomagnetic storm. See if your data allow you to say something like, "My data reveal that if the IGP delay is larger than (some value), then a geomagnetic storm is happening." If the data do not tell such a clear story, worry not: your project is not a failure. Instead, your data revealed that the IGP delay is not a useful way to predict geomagnetic storms.

(Oh, and asking these kinds of questions definitely doesn't count as getting someone else to do the work for you!)

Good luck with wrapping up your project! I'm done with travel until March 15th, so I'll be able to answer any questions you have between now and then much more quickly.

Hi, I'm starting my data analysis tomorrow and i want to make sure i don't miss anything. Does this sound like a good plan?:
Firstly, I wanted to ask if the data that i put into the graphs is an average of all the data. I was thinking that after i make the graphs that you mentioned, i could use the one that compares kp and igp to determine if igp delay can be used to predict kp values (this is one of the steps in the procedure). I'm basically recording the same data that I did in my experiment, but I will be using igp delay to predict kp values. I'm doing this for 3-4 days (I will record data six times a day, so, that's 18 - 24 predictions). Is 3-4 days too long?

I also have questions about percentages. I want to calculate percentages for:
- How accurate my predictions for kp values are
- How many times igp delay was higher when kp values were elevated
- How many times igp delay was higher when other space weather was occurring
- How many times igp delay was higher for unknown reasons
I want to calculate percentages for the last three things that i listed, but with lower igp delay. Should i put all the percentages i mentioned into a pie chart. Would i make separate pie charts for higher igp delay and lower igp delay?
Thanks for the help,
coolgirl285
p.s,
I'm entering my fair as an independent, so, my project should be done by the first week of March (this doesn't include the display) Is that doable? How long should i take to do my data analysis?
Again, sorry for asking so many questions.

Hi coolgirl285,

It sounds like you are on the right track with your analysis. I suggest putting all of your data on your graphs. Looking at trends of averaged data can be misleading. Yes, you can use the graph of igp delay vs. kp to determine if the igp delay can be used to predict kp. You could compare your predicted Kp values with the Kp values you measured. I don't think 3 to 4 days is unreasonable.

I'm not entirely sure that I'm following what the percentages you want to calculate. But, a pie chart is often a good choice for showing the fractions of things that add up to a whole. So, a pie chart sounds like a good fit for what you are doing!

I'm having trouble making the graph that shows how igp delay varied and the graph that shows how kp varied. I'm making the graphs on excel, so, if I put the time in column a, where do I put igp delay/kp?

Hi coolgirl285,

In Excel, try putting time in column A, then IGP delay in column B, and Kp in column C. If you select columns A and B and then do "Insert" -> "Scatter" -> "Scatter with Straight Lines and Markers" (the graph with dots connected by lines). This will graph IGP delay as a function of time. Then, select column A and C and do the same thing.

If you run into trouble or have other questions, just let me know and I will help you!

I'm not sure about my variables. I know my controlled variables are the times that i did my experiment, but I'm not sure about the independent and dependent variables since i have no control over igp delay or kp values.
I'm also confused about my data analysis. When i put the time in column a, would i do it like this:
8 am
11 am
2 pm
5 pm
8 pm
8 am
11 am
2 pm
5 pm
8 pm
etc.

Do make the graphs using the data for every single day i did my experiment? If possible, can you do an example of how i make the graph on excel because i really don't understand how to make the graphs.
thanks,
coolgirl285

When you are observing both the dependent and independent variable, it helps to go back to your hypothesis. You expect your measurement of "Thing I" will change based on the measurement of "Thing D"

Thing I depends on Thing D.

Example: Your measurement of "number of people who complain they are cold" depends on "the temperature of the room at the time of the survey."

More examples: http://nces.ed.gov/nceskids/help/user_g ... iables.asp

Hi coolgirl285,

I responded to one of your other posts about variables. HowardE and theborg gave some good advice. theborg suggested that Kp index is the independent variable and that IGP delay is the dependent variable. That's good advice. In my other post, I said that the independent variable was solar activity, which Kp measures. So, you could either list your independent variable as Kp or solar activity. Either choice would be fine. But the dependent variable is definitely IGP delay. And, yes, the time of day you measured the IGP delay was a controlled variables.

As far as graphs go:
When you enter the dates and times, in column A, enter them as follows. I'm making up dates here, so be sure to use the dates and times that you made your measurements! List all of the dates and times you collected data for in column A. Both the date and the time should go in the same cell. You will make one graph that shows the entire range of time you collected data.

4/8/15 8 am
4/8/15 11 am
4/8/15 2 pm
4/8/15 5 pm
4/8/15 8 pm
4/9/15 8 am
4/9/15 11 am
4/9/15 2 pm
4/9/15 5 pm
4/9/15 8 pm
4/10/15 8 am
4/10/15 11 am
4/10/15 2 pm
4/10/15 5 pm
4/10/15 8 pm
etc.

Excel will probably change the format of the values to something like 4/8/2015 8:00. That's OK. Excel might also change the times that are "pm" into 24 hour time (e.g., 4/8/15 8 pm might be formatted as 4/8/2015 20:00). That's OK, too. If Excel changes the dates and times to a slightly different format than either of these, that's OK, too.

Once you've filled in column A, put the Kp index you recorded for each of those times in column B. Put the IGP delay in column C. If, for example, the Kp delay was 1 and the IGP delay was 1.625 at 8am on 4/8/15 (again - I'm making these data up; these aren't the real values), then the row for 4/8/15 8 am would look something like this:

Column A Column B Column C
4/8/15 8 am 1 1.625

I've taken some screenshots of how to make the graph, which will hopefully be helpful to you. You can find them on this public webpage:

https://sites.google.com/site/scibudkpe ... creenshots

I have Excel 2010 on my computer, so the screenshots are from Excel 2010. If you're using a different version of Excel (e.g., 2007, 2013), then your screens will look slightly different than mine. But, the steps should be the same.

After you've entered all of your data, your spreadsheet should similar to the screenshot called "STEP 1 - DATA TABLE.PNG"

To actually insert the graph, select all of the data in all three columns and click insert -> scatter -> scatter with straight lines and markers. To select the data, click in cell A1 and drag down to the corner of the last cell with a number in it in column C. The screenshot called "STEP 2 - INSERT A SCATTERPLOT.PNG" shows you how to do this step.

After you click on the icon for "scatter with straight lines and markers", a graph should pop up that looks similar to the one called "STEP 3 - SCATTERPLOT.PNG".

The x-axis of the graph will probably look really weird, with numbers on top of each other. So, let's fix that. Start by selecting the x-axis. Then right click and select "Format Axis", as shown in the screenshot called "STEP 4 - SELECT X AXIS.PNG".

Look for where it says "Major Unit:". Change the radio button to "Fixed" instead of "Auto". Enter "1" in the box to the right of "Fixed". This will change the x-axis so that each day has only one label. If the x-axis is still too cramped, you can change the number to "7". This will make labels appear weekly instead of daily. The screenshot "STEP 5 - REFORMAT X AXIS.PNG " shows you how to do this.

Next, you will want to change the date format. Click the "number" option that is in the column of subtitles on the left side of the "Format Axis" window. Click on "Date" under "Category". Then choose one of the options under "Type". In the screenshots, I selected the one that says *3/14/2001. You could also use the one that says "3/14", if you prefer that date format.

Hopefully these steps will get you the graphs you need. Let me know if you run into more challenges.

Thank you to everyone who helped me. I won a bronze medal.
- coolgirl285

Hi coolgirl285,

I'm so happy for you! I'm glad things turned out well. You put in a lot of hard work and effort, and now you're seeing the rewards of all that hard work!