Leveraging public data to offer online inquiry opportunities

Abstract Inquiry activities have become increasingly common in Ecology and Evolution courses, but the rapid shift to remote instruction for many faculty members in response to the COVID‐19 pandemic has created new challenges for maintaining these student‐centered activities in a distance learning format. Moving forward, many instructors will be asked to create flexible course structures that allow for a mix of different teaching modalities and will be looking for resources to support student inquiry in both online and in‐person settings. Here, we propose the use of data‐driven inquiry activities as a flexible option for offering students experiences to build career‐relevant skills and learn fundamental ecological concepts. We share lessons learned from our experiences teaching a two‐semester course‐based research experience in global change ecology that leverages publicly available datasets to engage students in broadly relevant scientific inquiry.

• Introduction to the project • Organize jigsaw reading ( *Certain exercises may require more or less time than expected. We may need to change the laboratory schedule to accommodate this. If we do change the schedule, we will notify you in time for you to plan accordingly.

INTRODUCTION
Ecological questions require quantitative data. How researchers gather those data vary from project to project, but one of the important aspects of ecological research is figuring out how to quantify what is happening ecologically. Primary production is the conversion of solar energy into chemical energy stored in the bonds of reduced sugars and represents the base of the energy pyramid in an ecosystem. Facilitated by photosynthetic organisms, primary production can be quantified by measuring the accumulation of photosynthetic biomass over a period of time.
For this activity you will be exploring data on end of season biomass harvested from plots from the nearby field research station at Cedar Creek (and around the globe). Biomass measurements give an indication of the net primary productivity of an area (the amount of carbon fixed minus the amount of carbon used in respiration). The primary productivity of an ecosystem is dependent on the availability of factors that limit plant growth. Limiting factors for productivity can include sunlight, water, and nutrients. At the Cedar Creek station and at other sites around the world there are experiments set up to understand how nutrient input affects grassland ecosystems. The researchers conducting these experiments have joined together into a Nutrient Network or NutNet. You can read more about NutNet here (https://nutnet.umn.edu/). In short, this global experimental network it observing the effects of nutrient addition and herbivores on grasslands.

PART 1: GRAPHING DATA & USING A T-TEST TO COMPARE BIOMASS IN FERTILIZED & CONTROL PLOTS
Graphing baseline data: Using a t-test: In the initial year of a field experiment, it is important to verify that you have standard starting conditions across your field sites. In order to examine the impact of fertilizers on grassland productivity, we must first establish that all biomass production at each of the study plots is the same. To do this, we can use a statistical procedure called a t-test. The t-test allows us to compare two groups and determine if there is a statistically significant difference in the mean values of the two groups. To do this, we will use the data collected by students in fall 2018 from year zero in the Nutrient Network Education (NutNEd) plots to compare the living biomass measured in control plots and fertilized plots.

Procedure:
1. Open the data file labeled "Nut Net Bootcamp Data Set 0" this is the biomass data for the NutNEd plots at Cedar Creek that was collected by students in the Fall 2018 Foundations Lab. a. When you open the JMP program, choose File from the taskbar and then select Open. b. Navigate to the data file saved on your computer and click open, this should open the Excel Import Wizard. c. Leave all of the settings in the import wizard at their default values and click import. This will open your file as a JMP table. Your table should contain 4 columns: "site_name, year_trt, trt, and live_mass" 2. To make things a bit easier for you, we've already subsetted this dataset so that it only includes the biomass measurements for the control plots and NPK plots from year zero in NutNEd. We will perform the t-test to see if there are any differences between the control and treatment plots in our year zero samples. A t-test is a statistical test used to test for differences in the means of two groups. In this case, we are looking for evidence that there is no difference between our two groups. a. Under the analyze menu, select the fit Y by X option. This will launch the fit Y by X window. b. Select the trt variable as your X factor and the live_mass as your Y response variable. Click okay to run the analysis. This will open a window containing a plot of your data.
c. In the upper right-hand corner of your plot window, there is a small red triangle that allows you to run various statistical tests. Click the arrow and select t-test from the dropdown menu. d. The t-test analysis table should open below your plot. There are several things listed in this table, but the thing that we are looking for is called the p value. This value is located next to the label that says "Prob<t". It is common practice for scientists to use a p value of 0.05 as their cutoff value for determining statistical significance. In other words, if your p value of a t test is greater than 0.05 there is not enough evidence for you to conclude that there is a true difference between your two groups (i.e. that any observed difference could be due to chance). Look at the p value listed in your table. What conclusion would you draw about the difference in live biomass between the control and treatment plots in year zero?

PART 2: GRAPHING PRIMARY PRODUCTION ACROSS TIME AND SPACE Graphing Primary Production Across Time:
For the first part of this lab, we will explore how primary production changes over time in response to a nutrient addition experiment. You have been provided data from the Nutrient Network, a global network of ecologists performing standardized experiments to better understand how fertilization impacts diversity and productivity in grassland ecosystems. Using this dataset, you will use a simple linear model to answer the following questions: 1) How has primary production changed over time in grassland plots that have been fertilized with NPK 2) How does this pattern compare to what happened in the control plots over time?

Procedure:
Using the data from the Cedar Creek NutNet plots, you will work with a partner to graph the annual primary production values for each year of the experiment in both the control plots and the fully fertilized plots. You will use a simple linear regression to determine if productivity in these two plots has increased over time and compare the rate of increase in the fertilized plots to the rate of increase in the control plots. a. Navigate to the graph tab on the toolbar and open the graph builder b. In the graph builder window, drag the live_mass variable to the y-axis of your graph c. Next, drag the year_trt variable to your x-axis d. Finally, you want separate regression lines for the control and NPK plots. To achieve this, drag the trt variable into the overlay box. This should separate your graph into red dots and blue dots representing the two treatment types. e. Next, you will need to change the graph type to linear regression by selecting the straight line on a scatter plot icon at the top of the graph builder window. Once you have changed the graph type, you can click done in the graph builder window. f. In this window, you can click on the legend, axes, and other aspects of the graph to customize the visualization. Modify your figure by providing appropriate axis labels, legend labels, etc. Once you have a figure you are happy with, save your figure by going to file, export, and exporting the figure as an image file. You will upload this image file a part of your post-lab assignment. 3. For the last step in this activity, we will use the analyze tab to test the significance of slopes for each treatment group. In other words, we want to determine if there is statistical evidence that the slopes of our two lines are significantly different from zero (flat). Then we will compare our slope parameters to determine if our fertilized plots have a larger slope value than the control plots (a faster rate of biomass growth over time).
There is a statistical test to determine is this difference in slopes is significant called an ANCOVA (Analysis of Covariance), but this test is beyond the scope of this activity. a. Start by opening the Fit Y by X program under the Analyze tab in the toolbar. This will open the Y by X model window. b. Add year_trt as the x factor and live_mass as the y response. This will result in a simple linear model relating these two variables. We want to see if there is a difference between the control plots and treatments, so add the trt variable to the By option. Hit okay to run the analysis. c. After you hit okay, the Bivariate relationship window will be opened. This should contain two scatter plots, one for the control plots and one for the NPK plots. d. In the upper left corner of the bivariate plots, you should see a small red triangle.
Click the triangle to launch the fit options. Click the fit line option to produce the best first line. e. The best fit line should now be shown on the bivariate plot and the linear fit window will appear under the plot. This window contains a lot of information, but for our purposes we are interested in the Analysis of Variance tab and the Parameter Estimates tab. f. The Analysis of Variance tab tells you if the slope of your line is significantly different from zero. In the bottom right corner, there is a section labeled Prob>F, this is referred to as the P value for the statistical test. If the P value is less than 0.05, it means that there is statistical evidence that the slope of your line is not zero. Record the P value for your control plots. How would you interpret your P value? g. Under the parameter estimates, you will see the slope and intercept estimates for the simple linear model describing your best fit line. The slope estimate is found in the year_trt row of the table. Record the estimate for your slope parameter. How would you interpret your slope parameter? h. Repeat the above steps on the bivariate plot window for the NPK plots to find the P value and slope estimates for the NPK plots.
i. Now compare the slope estimates and P values from the two treatment groups.
What can you conclude about the effect of the NPK treatment on biomass production?

Graphing Primary Production Across Space:
In addition to understanding how productivity changes over time, the Nutrient Network experiment allows us to study how fertilization impacts primary productivity at different grassland sites across the planet.

Procedure:
To start, we will compare the live biomass in fertilized plots and control plots at the original Nutrient Network sites that have now been running for 10 years. Then we will explore how this fertilization effect varies by latitude. a. Open the graph builder tool. Drag the site_name variable to the X axis and the live_mass to the Y axis. b. To show the difference between the control and treatment plots, drag the trt variable to the overlay box. c. This figure can be displayed as a bar graph or a box and whisker plot. These options are found on the icon bar above the plot. Experiment with the two types of plots and choose the plot that you feel is the best way to represent this data. d. Once you have picked your plot type, click done and label your axes to finish your graph. You will upload this figure on Canvas as part of your post-lab assignment. Comparing the Impact of Different Fertilizers:

Procedure:
Using the data from multiple NutNet plots, you will work with a partner to graph the annual primary production values for each year of the experiment in both the control plots and the fertilized plots. You will use a simple linear regression to determine if productivity in these two plots has increased over time and compare the rate of increase in the different fertilized plots to the rate of increase in the control plots. a. Navigate to the graph tab on the toolbar and open the graph builder b. In the graph builder window, drag the live_mass variable to the y-axis of your graph c. Next, drag the year_trt variable to your x-axis d. Additionally, you want separate regression lines for each treatment type. To achieve this, drag the trt variable into the overlay box. This should separate your graph into different colors based on the treatment type. e. Finally, we want to determine if the relationship for each treatment type is the same for each of the sites we are looking at. To do this, you can use the "wrap" function in JMP. Drag the site_name variable in to the "wrap" box that is in the top right corner of the graphing window (just to the left of the overlay box). f. Next, you will need to change the graph type to linear regression by selecting the straight line on a scatter plot icon at the top of the graph builder window. Once you have changed the graph type, you can click done in the graph builder window.
i. Compare and contrast the regression lines from each of the different plots. Does it appear the nutrient treatments had the same effect at each site? Would you feel confident saying there was a single nutrient (N, P, or K) that limited primary production at all sites? g. In this window, you can click on the legend, axes, and other aspects of the graph to customize the visualization. Modify your figure by providing appropriate axis labels, legend labels, etc. Once you have a figure you are happy with, save your figure by going to file, export, and exporting the figure as an image. You will upload this figure as part of your post-lab assignment. 3. For the last step in this activity, we will use the analyze tab to test the significance of slopes for each treatment group. In other words, we want to determine if there is statistical evidence that the slopes of our lines are significantly different from zero (flat). Then we will compare our slope parameters to determine which fertilized plots has the largest slope value compared to the control plots (a faster rate of biomass growth over time). There is a statistical test to determine if this difference in slopes is significant called an ANCOVA (Analysis of Covariance), but this test is beyond the scope of this activity. a. Start by opening the Fit Y by X program under the Analyze tab in the toolbar. This will open the Y by X model window. b. Add year_trt as the x factor and live_mass as the y response. This will result in a simple linear model relating these two variables. We want to see if there is a difference between the control plots and treatments as well as if those relationships were the same at each site, so add the trt and site_name variables to the By option. Hit okay to run the analysis. c. After you hit okay, the Bivariate relationship window will be opened. This should contain 16 scatter plots, one for each treatment at each site.
d. In the upper left corner of the bivariate plots, you should see a small red triangle. Click the triangle to launch the fit options. Click the fit line option to produce the best first line. e. The best fit line should now be shown on the bivariate plot and the linear fit window will appear under the plot. This window contains a lot of information, but for our purposes we are interested in the Analysis of Variance tab and the Parameter Estimates tab. f. The Analysis of Variance tab tells you if the slope of your line is significantly different from zero. In the bottom right corner, there is a section labeled Prob>F, this is referred to as the P value for the statistical test. If the P value is less than 0.05, it means that there is statistical evidence that the slope of your line is not zero. g. Under the parameter estimates, you will see the slope and intercept estimates for the simple linear model describing your best fit line. The slope estimate is found in the year_trt row of the The pre-and post-lab questions for this lab exercise are two separate Canvas assignments.

PRE-LAB QUESTIONS
1. List the first 10 words that come to mind when you think of "grassland". 2. Browse https://nutnet.umn.edu/ What anthropogenic activities have altered nitrogen and phosphorus availability? How much have their availabilities increased? 3. Give a brief description of the NutNet project in your own words.

POST-LAB ASSIGNMENT
This assignment consists of two parts, which you can upload to Canvas as separate files, or combine into one Word or pdf document: 1. Upload a document that contains all of the plots you have made from this lab activity (the ones indicated as being part of the post-lab in bold font in this document). 2. Upload a document with your answers to the post-lab questions.

POST-LAB QUESTIONS
1. Without looking at your first list, do a second quick word list of 10 words that come to mind when you think of "grassland". 2. Describe the pattern you observed when you graphed primary production across time for the Cedar Creek NutNet plots. Where the fertilized plots different than the control plots? Explain your answer. 3. How did the fertilization effect of the NutNet plots vary across space (latitude)? What conclusions did you draw based on the graph you made? 4. Data analysis is an important part of doing science and as scientists we often have to choose the correct tools for performing our data analysis. In this lab, you used JMP software to do your data analysis. Note: Attendance at ONE of the poster sessions in week 13 is required. Dates and times for these sessions will be announced early in the semester. Your entire group doesn't need to attend the same poster session together. If you're unable to attend any of the sessions, please inform your instructor as soon as possible.

Assignment 1 Linear Regression and Correlation Testing Linear Regression
Simple linear regression describes a continuous response variable (prediction) as a function of one or more predictor variables. Remember high school math? The following is a linear model you would have learned in high school: = + In a linear model, the slope (m) describes how much of an effect x has on y and b represents the intercept value (the y value when x is zero). Let's talk about the price of a small bag of rice. If I have a linear model of rice price, y, as a function of rice quantity, x, and the slope is equal to 0.00001 (m=0. 00001), then each increase in a grain of rice would result in $0.00001 increase in the price of the bag of rice. So the linear model of the rice price would be: = 0.00001 + The parameter b tells us how much a bag of rice would be worth if it had no rice inside, so b would the cost of the bag itself. We can imagine the bag costs $0.50, so: = 0.00001 + 0.50

What if we want a model to understand what leads to weight gain? Our steps would be:
Step 1: Define a dependent variable (what are we using to measure weight gain). How about body fat percentage? That would be variable y.
Step 2: Pick a variable that might influence weight gain. How about number of burgers eaten per week? This is the independent variable, x.
Step 3: Gather data on body fat percentage and burgers eaten per week.
Step 4: Create a regression model. In this example, body fat percentage is the dependent variable (y) and burgers eaten per week is the independent variable (x). Other influences, like starting body fat percentage, amount of vegetables eaten per week and amount of exercise per week all get combined into another parameter β that we will mention below. A plot of that data could look something like this: Note: The data points don't fall on the line. Why? The line is a best fit that minimizes the total distance of all the data points to a straight line. It tells us that, on average, the more burgers you eat per week, the higher we can predict your body fat percentage to be. The distance from a data point to the line looks like noise or error, but it actually captures all of the other influences like exercise and vegetable intake. To include these other variables we can create a multiple variable linear equation of the form: = 0 + 1 1 + 2 2 +. . . ..

Linear Regression in JMP
In JMP, you can perform a linear regression using the analyze function. Think back to your bootcamp activity where you examined how the amount of primary production changed over time in fertilized and control plots. This was a linear regression. To build a linear model in JMP, select analyze from the menu bar and then choose "fit Y by X". From here, you can define you x and y variables that you would like to build a regression model from.
For this assignment, you will build a linear regression model to describe the relationship between added nitrogen and aboveground productivity. You can access the data from the Long Term Ecological Data network data portal. Click HERE.
The get the data you must: 1. Download the data file through clicking the above link 2. If the filename is e247_Aboveground Standing Crop Biomass.txt.csv, change the file name to e247_Aboveground Standing Crop Biomass.txt 3. Open the file with JMP Using this dataset, we will build our linear regression model: Step 1: You will need to determine which will be your dependent variable and which will be your independent variable. The variables to choose are: • Nitrogen fertilizer added to plot • Aboveground biomass Consider the experimental design of the nutrient network in making your decision.
Step 2: Build your linear regression model using the Fit Y by X function under the analyze tab. Drag your dependent variable into the Y box and your independent variable into the X box. Hit okay to produce the scatterplot. Fit your linear regression model by clicking on the small red arrow in the top left corner of the scatterplot window and selecting "Fit Line" From the linear fit model, record the following parameters 1. Slope (m): 2. Intercept (b): 3. R2 value: 4. t-Ratio: 5. p-value: In addition to the parameters above, you can find information about a number of other statistical values in this output table. Below are some short descriptions of the different values you can find (we will spend more time on this in future assignments).
Residuals tells us the variance that is not explained by the independent variable. It is the distance of the points from the best-fit line. It tells us the Min, Max, Median and interquartile range (1Q, 3Q). If the range is roughly symmetrical (in terms of magnitude), then the data is likely normally distributed.
The Estimate values are the estimates of the mean of the distribution of the variable listed (intercept and slope) The standard error (Std. Error) is the square root of the variance of the distribution. It is a measure of the uncertainty in the estimate.
t-value is the estimate divided by their standard errors.
Pr(>|t|) is the probability of achieving a value of t that is as large or larger, if the null hypothesis were true. Here the null hypothesis is that the estimates are individually 0 (intercept is 0 and there is no slope (the line is horizontal)).
The significance codes give a breakdown of what the stars (**) mean, and thus explain the p-value significance.
The Residual standard error tells us how much variability there is in the residuals.
The R-squared tells us information regarding the correlation between the dependent and independent variables. If the ratio is close to 1, there would be a 'perfect' correlation. If the value is 0 there is no correlation. The Adjusted R squared accounts for the number of variables in the model.
The F-Statistic tells us the significance of the entire model. It is the ratio of two variances, the variance explained by the parameters in the model (in this example we only have one) and the residual or unexplained variance. The bigger it grows, the more unlikely it is that the parameters in the model do not have any effect at all.

Appendix 5: Research paper guidelines (used in multiple research areas)
Biol 3004 Research Paper: Guidelines for Writing

Overview
In this assignment, you will be writing a scientific research paper based on your lab work this semester. Scientists write for publication in peer-reviewed journals because this is how we share our results, analyses and insights with the broader scientific community. Research papers are one of our most important products, and a project is not complete until it is published.
Writing for publication in peer-reviewed journals serves a very different function than writing a lab report. Lab reports test your understanding and ability to communicate aspects of your experiment to an instructor; generally these experiments have been repeated multiple times and have an expected result, interpretation or outcome. In contrast, a research paper aims to add to the body of knowledge within the scientific community. Research papers use data from experiments and analyses of those data to support the authors' conclusions. Other scientists who read a research paper may have their own interpretations of the data and may not agree with the authors' conclusions. They may want to repeat the experiments, or design new experiments to test alternative hypotheses that may lead to different conclusions.
Generally, research papers are not written like an essay from start to finish. Usually scientists have already performed most of the experiments, analyzed their data and drawn conclusions from the data before starting to write. Therefore, it's common to begin by creating draft versions of the expected figures and tables that will document the results obtained. The writing itself may begin with the Results section, using the assembled figures and tables (and revising them if needed during the writing). Alternatively, writing may start with the Materials and Methods (since it's often a straightforward section to write and that makes starting the writing process a little easier), or with the Introduction (to lay out the background for the study). The Discussion and Abstract are often written last, after the other sections are complete.
For this course, you'll be writing the introduction first so you have more time for project work before trying to write the other sections. Due dates for drafts are outlined on the Canvas site. For each section of the paper, you'll first submit a draft for peer review. You'll review drafts written by two other students, and receive feedback from students who read your draft. You can revise your initial draft (if you wish) before submitting it to your grad TA through a separate Turnitin assignment on Canvas. You'll receive feedback from your grad TA on each draft and have the opportunity to make revisions again before submitting the final version of your paper at the end of the semester. The final version of your paper is due by 11:55PM on Wednesday May 6. Allow time for editing and proofreading before you submit your final revised draft!

Purpose
This writing assignment has a number of learning objectives: 1. To give you an authentic experience of communicating experimental results in a typical format for biology.
2. To challenge you to draw specific conclusions from your data and to defend these conclusions. 3. To challenge you to integrate information from a variety of sources to construct and support your conclusions. 4. To give you the opportunity to improve your writing through revision based on critiques from your classmates and instructors.

Format & Content
Your paper should contain the sections listed below. It should be double-spaced, in 12-point font with 1-inch margins. There is no required number of pages; suggested lengths for some sections are listed below, but these are only rough guidelines. Previous students who've received high scores on their research papers have written at least 8 pages (not including references, figures, and figure legends). Papers more than 20 pages long should be edited to reduce their length.
In this course, project work is carried out in groups, but each student is responsible for writing their own research paper. The figures in your paper may be identical to those in the papers written by your team members, but the text should be your own. Everyone in the group is expected to contribute to making the figures for their research project.

Title page
The title should indicate what the paper is about and be interesting enough to encourage someone to read the paper. Don't forget to include your name on the title page!

Abstract (~150-250 words)
The abstract should provide a concise summary of the major aspects of the entire paper.
• Purpose of the study (hypothesis, overall question, objective) • Description of the basic experimental design/study design • Major findings/results • Important conclusions or interpretations

Style
• Past tense • Should stand on its own without reference to any other part of the paper • Limit background information to a sentence or two, at most.
• Must be consistent with what is reported in the rest of the paper

Introduction (~2 pages, double-spaced)
• The introduction should provide context for the work, the objectives, and the hypothesis/es. • What is the broad question that motivated this study? Why is the problem interesting? Establish the context for the work by summarizing our current understanding of the problem being investigated, with references to the relevant primary research literature, so the reader can understand how the study relates to previously published work.
• State the purpose of the work in the form of the hypothesis, problem, or specific question(s) it addresses. How does the work fill a gap in our understanding of this area of biology? • Briefly explain the approach to answering the specific question(s) and the overall strategy for the study.

Style
• Use the active voice as much as possible. Some use of first person is acceptable.
• Cite primary scientific literature (see Literature Cited below) to document any information that is not common knowledge.

Materials and Methods (2-3 pages, double-spaced)
The Materials and Methods should document the necessary materials and the experimental procedures in sufficient detail that an experienced scientist could reproduce the work.
• What strains or biological starting materials were used for the experiments? How were they grown or maintained? • What other materials did you use? Provide sources for important or unique reagents that were critical for the work.

Style
• Usually written in 3rd person, passive voice. Use the past tense.
• Divide this section into logical subsections devoted to specific procedures or groups of procedures; each subsection should have its own sub-heading. • Write in paragraphs and full sentences, not a list of bullet points. • Cite earlier papers or manufacturer's instructions for procedures that have been published previously. If you cite a previous publication you don't need to report details of the procedure, but you should describe any differences between your procedure and the published one.

Results (1-2 pages, double-spaced, not including tables and figures)
The purpose of the Results is to present and illustrate the research findings.
• Present results in an orderly and logical sequence so the reader can follow the scientific "story". The order in which results are presented doesn't need to match the order in which they were obtained. • Use both text and figures/tables to present your results. Data generally should be shown in figures or tables, though some may be presented only within the text. • In the Results text, briefly provide context for the data by describing the question addressed or the approach taken. Avoid redundancy by omitting procedural detailsleave them for the Materials and Methods or the figure legends. Link elements of the story to one another with explicit logical connections between paragraphs and subsections. • Describe important trends or point out key observations in the data shown in each figure or table. Descriptions may include objective first-order interpretation or conclusion(s) from the data, but more extensive interpretation should be saved for the Discussion. Reference each figure or table as you describe it (" Fig. 1", "Fig. 2", " Table  1", " Fig. 3", etc). • Address unsuccessful experiments appropriately. If you had a failed attempt, but were subsequently successful, you don't need to mention the failed attempt. If an experiment didn't go as expected but you learned something from the attempt, focus on what you learned and describe any limitations on the information obtained. If an experiment was unsuccessful, state that fact and describe what you did instead and/or subsequently. Don't include lengthy descriptions of the possible reasons for failure. • Keep in mind that it's OK to have negative results; these should still be presented and their meaning explained.

Style
• The Results section is usually divided into subsections with sub-headings to clarify the flow of the story. • Use the past tense to describe what you did. Use active voice as much as possible.
• Each figure and table should be sufficiently complete to stand on its own. Every figure or table must have a legend, or caption, describing its contents. Legends need to be sufficiently detailed that a reader can understand exactly what's in the figure or table without referring to the corresponding Results text. Legends should briefly describe how the data were collected as well as what is shown in the figure or table itself. Make legends single-spaced or in smaller font to distinguish them from the main text. • Figures and tables may be integrated into the text, or they can all be placed at the end of the paper. Number each figure according to the order in which it's discussed in the text, and present them in that same order. Number tables similarly, but with their own set of numbers. If a figure consists of several parts, each part should be labeled with a different letter. • Be specific, accurate and precise when describing your data. Relate the results from experimental samples to negative and positive controls as appropriate.

Discussion (~2 pages, double-spaced)
The goal of the Discussion is to provide an interpretation of your results and support for your conclusions, using evidence from your experiments and the results from prior studies. The Discussion should also describe the impact of the findings.
• Briefly summarize the findings in relation to the broad question posed in the Introduction. Explain whether your hypothesis is supported or rejected. Bring the paper full circle by connecting the new information back to the big picture from the beginning. • Explain your observations as much as possible, focusing on mechanisms. If there are reasonable alternative explanations, state them and contrast them with your own interpretation.
• What specific conclusions can you draw from your experiments? Explain how your results agree or disagree with prior research and how they add to the existing body of knowledge. • How would you move forward with this research based on the information you obtained? Provide suggestions as to what should be done next and/or describe new questions your study raises.

Style
• Refer to work done by specific individuals (including yourself) in the past tense.
• Use the active voice as much as possible. Some use of first person is acceptable.

Acknowledgements
This section recognizes those who provided assistance with the study and/or necessary materials.
• List people who contributed to the work presented in the paper, such as group members who helped with the project, mentors who provided guidance, and anyone else who made a significant contribution to the paper.

Literature Cited
This section lists the references cited in the body of the paper. This section is extremely important for giving fair credit to the work of previous researchers, presenting your findings in the context of what has already been done, and avoiding accusations of plagiarism.
• Cite primary scientific literature (books, journal articles, review articles) whenever possible. Provide full reference information for the sources you cited in your paper, in the style of the journal Cell (see below). • Be wary about the quality of online information (eg: .gov websites are more authoritative than .com or .org sites). If you cite a website, give the author and/or title, date of most recent update, URL and accession date. • If you obtained information from an individual based on his or her own experiments or experience, you may cite that person as a source (eg: J. Doe, personal communication). Before citing a personal communication in paper being submitted for publication, it's customary to get permission from the person being cited. You don't need to observe that formality in this course, but you should check if the information is published before citing a person as a source.

Resources for Writing
• "A Short Guide to Writing About Biology" (9th ed. or 8th ed.) by Jan Pechenik

Grading
Research papers will be evaluated using the accompanying research paper rubric. Review the specific criteria for each section and refer to them while writing and editing your paper. Final papers will be graded out of 200 points.
Participating in the peer review process will earn up to 10 points each time: • 5 points for submitting a draft for peer review • 5 points for providing feedback on two peer drafts Submissions for peer review and the feedback provided to other students will be monitored for quality. We reserve the right to withhold credit for minimally completed drafts or poor-quality feedback.
The drafts of each major section (Introduction, Materials & Methods, Results, and Discussion) that you submit for grading by your grad TA will be scored out of 10 points. A draft of the Abstract is due along with the draft of the Discussion, and will earn up to 5 additional points.

BIOL 3004 Poster Guidelines Overview
The purpose of these posters is to communicate the results of your independent research projects to your fellow BIOL 3004 students, to BIOL 1961 students, and to other interested scientists from outside the course. Poster sessions will be held from April 20-24. Making the poster is a group assignment. Every group member is expected to attend one of the poster sessions to present their group's poster, either alone or together with other members of their group. In addition, each student is expected to review at least three posters produced by other groups in the course. • Submit poster reviews online by 11:55PM on Sunday April 26

Poster format
The maximum final poster size is 36 inches high x 44 inches wide (in landscape orientation). It's probably easiest to make your poster as a single large Powerpoint slide, but you can use any software that all members of your group are comfortable using. The final files you submit for printing and to Canvas must be in PDF format.
There is lots of advice about making effective posters available online; some links and recommendations for format and design are given below. If desired, you can download a poster template as a starting point for your poster. If you do, be sure to adjust the dimensions of the poster to the size above before starting to work on your design.
Posters are a visual medium for communicating your work to potentially interested people, with an opportunity for back-and-forth interaction between you and your audience. Ideally, posters should be viewable from a distance to give people an idea of what your work is about and what you've found, in a way that encourages them to come closer and talk with you to learn more about it. At some scientific meetings posters are displayed continuously, even when the author isn't present; thus it's also useful to design posters so a viewer who doesn't have the chance to talk with you can look at your results and understand your project.

Recommendations:
• Limit the amount of text on your poster to ~600 words or less. Use visuals (with informative labels) to convey your message when you can. Use bullet points rather than paragraphs, though short paragraphs may be effective in some situations. • Use large font sizes for text, including labels on figures. To ensure your poster can be read from a distance, print out a test version on an 8.5x11 sheet of paper and check that you can read all the text. Pay special attention to axis labels on graphs and to the text in tables and legends: all text should be legible from a distance. • Show your test print to a friend to check the overall flow and visual appeal of the poster.
Specific tips from the University Imaging Centers (where your poster will be printed): • Fonts: If you submit a file with a non-system font (meaning one you downloaded from the internet), you must either supply us with the font, or convert your file to PDF. • Certain effects in PowerPoint, such as transparencies and gradients, do not translate properly to print. If you are using such effects we recommend printing a small proof for review. [Note: Printing a small version on your own printer may give you an idea about whether a specific effect will work or not, but results with the UIC printer may be different. It's best to avoid these effects rather than have your final poster turn out looking odd.] • When inserting charts, figures, and photos into your PowerPoint file, please use either the "Insert" menu, or the "Paste Special" command. Do not do a simple cut and paste, otherwise there may be printing errors. • If you're working with an Illustrator or InDesign file, please make sure all images are embedded, and that your text has been converted to outlines. [Do this before making a PDF, then check that the PDF looks OK before you submit it.] • Prints will have a 1/4 inch white border outside of the printable area. • Make sure that there is at least once inch of space between the page border and any text, photos, or other objects.

Poster content
Your poster should contain the sections listed below. You can include other elements as needed to describe your work. Your poster does not need to cover every aspect of your project: pick the key results (or a single result) and focus on communicating those key points effectively. You can elaborate further on your work when you present the poster, if some people want to hear more details or additional results.

Poster printing
Upload your final poster in PDF format to the University Imaging Centers printing site (http://uic.umn.edu/node/add/uic-poster-submission) by the submission deadline. Your file must not be larger than 20MB. In the Contact section, enter your own contact information, with Catherine Kirkpatrick in the PI Name spot, Biol 3004 as the class, and CBS as the college. Enter your group number, including a letter or two to indicate your research area: C for computational microbiology, E for microbial experimental evolution, GE for global change ecology, M for zebrafish microbiome, or Z for zebrafish environmental toxicology. You may choose to pick up your poster in 23 Snyder Hall (St Paul), 1-151 Jackson Hall, or 1-220 Cancer & Cardiovascular Research Building (both Minneapolis East Bank).

Poster sessions & poster reviews
Students are responsible for bringing their group's poster to each session where a group member is scheduled to present it.
Each poster session will be divided into two halves of about 45 minutes each. Unless multiple people from the same group are presenting at the same session, each poster typically will be presented in either the first or second half of a session. Poster sessions will be a bit longer than 1.5 hours to allow some time for taking down posters at the end of the first half and putting up the posters for the second half.
Every student will be expected to review at least three posters produced by other groups in the course. One of these posters will be assigned to you (to ensure that every group gets a minimum number of reviews); for your other reviews you can pick any posters you wish. You'll need to listen to a presentation about each poster you're reviewing so you can evaluate the presentation as well as the poster itself. Generally students will be assigned to evaluate a poster in the other half of the same session where they're scheduled to present, and most students complete their other reviews during the same session. You're welcome to attend other poster sessions to complete additional reviews if you wish.
Reviews will be submitted through an online form, and are due by 11:55PM on Sunday April 26. Printed copies of the questions will be available during the regular poster sessions for those who want to take notes about a poster and submit their review later, but paper reviews will not be accepted for credit. You'll earn 8 points for each of your 3 reviews, for a total of 24 points. You can earn extra credit by attending an additional poster session (other than the one at which you are presenting) and submitting up to two additional reviews (for 2 points each).

Grading
In addition to submitting your final poster PDF for printing, you must also submit it through the "Final poster" assignment dropbox on Canvas. Your grad TA will evaluate your poster for completeness using the criteria under Poster content above, as well as for overall effectiveness and organization. Posters will receive a final grade out of 100 points based on 75 points from your grad TA and 25 points from poster reviews (see below). A research question or hypothesis is hard to identify, or very little background is provided. The utility of the results or value of the project is unclear. 0-6 pts.

Approach & methodology 10 pts
Is it easy to understand how the data were gathered and how the analyses were performed?
The steps of the project (data collection and/or analysis) are well explained or diagrammed. 9-10 pts.
Descriptions of the study design and methodology are mostly clear and accurate, but viewers might have questions about some aspect of the work. 7-8 pts.
Descriptions of the approach, methodology and/or analysis are missing or hard to understand. 0-6 pts.

Strong Satisfactory Weak
Data figures and tables 10 pts.
Are data graphed or presented in an appropriate format? Are the data figures easy to interpret?
Figures, tables and legends are very easy to read and interpret; they meet the "Acid Test" criteria of the Sample Figure  & Legend assignment. 9-10 pts.
Generally appropriate choices of presentation format, but some data may be graphed in a format that doesn't support the intended interpretation. Visuals are labeled well enough to read easily and have explanatory legends. 7-8 pts.
Some figures are difficult to interpret due to defects in labeling and/or legends, or data presentation is significantly flawed or misleading. Labeling and legends are insufficient for the reader to understand the results. 0-6 pts.

Use of figures, images and diagrams 10 pts.
Are the visuals useful and relevant to the overall message of the poster?
Appropriate number and types of graphics are used to convey the work and support the take-home message. Visuals are attractive, large enough to see details, and easy to understand. 9-10 pts.
One or two of the visuals are not relevant, not large enough, or hard to understand, but the poster take-home message is still apparent. 7-8 pts.
Insufficient use of visuals diminishes the impact of the poster, or the visuals distract from, rather than helping to convey, the overall message. 0-6 pts.

Results and conclusions 10 pts
Does the poster clearly state the main findings of the study? Are the conclusions supported by the data?
The results and conclusions are clearly but briefly stated in text. Data analyses are free of errors. The conclusions represent a reasonable interpretation of the data. 9-10 pts.
Data analyses are fairly complete but may have minor flaws; conclusions are stated, and are mostly consistent with the data shown. 7-8 pts.
Data are presented in figures with no statement of results or conclusions, or data analysis is flawed so that some conclusions are not supported by the data shown. 0-6 pts. Is the poster well organized and easy to read, with an effective layout? Is the text sufficient to explain the graphics, but not overwhelming to the reader?
Individual parts of the poster combine to form an easily navigated unit; spacing and flow allow the viewer to follow the story readily. Graphics combine well with text (that is brief and easy to read) to deliver the message. 9-10 pts. Oral presentation of poster (10 pts)

Mastery of topic; ability to answer questions 5 pts
Did the presenter explain the project effectively?
The presenter highlighted the main points, explained the approach and results, and was able to answer questions about the work. 4-5 pts.
The presenter did a fair job explaining the project and was able to answer some questions about the work. 2-3 pts.
The presenter struggled to explain or answer questions about the work, and appeared not to understand some parts of it. 0-1 pt.

Engaging explanation of poster 5 pts
Did the presenter guide you through the poster in an enjoyable way?
The presenter made the work interesting and responded to their audience. The presenter spoke clearly and was easy to follow. 4-5 pts.
The presenter explained the project but didn't engage much with their audience or was somewhat hard to follow. 2-3 pts.
The presenter read the poster rather than explaining it, or they were hard to understand and/or didn't interact with their audience. 0-1 pt.
Please comment on the poster, including which features most helped you understand the research project and what was most confusing or difficult to understand.

Introduction
The effects of global climate change have become the subject of intense public and scientific concern within the 21 st century. As the climate changes the atmosphere is expected to have an average increase in heat of at least 2.7 °F over the next 70 years (Collins et al., 2013). One of the areas that has already been rapidly changing due to this heating are regions of arctic tundra.
Tundra regions have already been observed to decrease by 30 centimeters a year (Walker, 2007).
As this loss of permafrost in the tundra occurs a startling relationship between climate change and permafrost reveals itself. One of the main agents of global heating is the greenhouse effect caused by a variety of gases in the atmosphere. One gas, CO2 has been the subject of interest due to its radical increase in the atmosphere due to human activities (Dlugokencky et al., 2019). With the relationship between CO2 and climate change being apparent, a related relationship with permafrost mentioned above is also revealed. Regions of tundra have acted as large carbon sinks for at least the last 2,000 years (Billings et al., 1982). As more permafrost melts due to heating, more CO2 is expected to be released into the atmosphere leading to further heating and potentially a runaway feedback loop of warming that could devastate the ecology of the tundra (Oechel et al., 1993).
There has been much research done on the tundra such as Romanovsky, 2001, Ma et al., 2019, and Billings et al., 1982. Numerous measurements of Gross primary production (GPP), net ecosystem production (NEP), carbon flux, organic matter content, net ecosystem exchange (NEE) and many other factors have been made throughout many regions with permafrost by the Arctic Data Center. This existing research has indicated that the melting of permafrost is increasing and with that the rate of ecosystem exchange is also changing (Segal et al., 2016). The exchange of carbon and other greenhouse gasses such as methane in tundra has increased throughout the 20 th and 21 st century and is expected to continue to increase (Marrero, 2010).
One thing that is often overlooked when studying changes in regions of tundra is the fact that there exist many types or zones of permafrost, continuous, discontinuous, sporadic, and isolated (Lantuit, 2016). Each type of permafrost exhibits unique qualities that may change how they respond to the changing climate. For example, discontinuous permafrost exhibits melting from both top and bottom unlike continuous permafrost (Romanovsky, 2001) There has been previous research on how organic matter and matter densities differ between the different zones such as in Ma et al., 2019. While there is data and research on how climate warming will affect tundra and permafrost as a whole, we do not know how this heating will affect each zone of permafrost independently and how these differences are significant. Insights in to how each of the different zones will function as carbon sinks or sources and how they will interact with the future climate could be gained from research into how these zones are affected. Of the zones, continuous and discontinuous permafrost are of particular interest due to their high abundance and regular distribution throughout the arctic (Lantuit, 2016).
This study has the goal to understand how the NEE has changed in the last decade in permafrost, specifically in continuous and discontinuous permafrost zones. We utilized eight datasets from across the northern hemisphere to analyze the trends in NEE using longitudinal and temporal data. Linear regressions of each permafrost zone were plotted in order to understand the temporal relationship with NEE in each zone and then an ANOVA was conducted to analyze the relationship between the changes in the different zones. This allows us to ascertain if there is a significant difference in the effects of climate change on different zones of permafrost, what degree this difference is, and if one zone should be given priority to preserve.
We aimed to answer the question of how soil carbon exchange into the atmosphere has changed in continuous and discontinuous permafrost. We hypothesized that both discontinuous and continuous permafrost zones will show positive net ecosystem exchange values over time. In comparison to continuous zones, discontinuous zones will release more carbon due to more plentiful ice cracks that promote thawing; also, discontinuous zones will respond to changes in climate more dramatically due sunlight penetration through ice cracks which warms deeper ice layers.

Study Sites
NEE data was compiled from five different sites, three of which were in continuous permafrost zones and two of which were in discontinuous permafrost zones. The continuous permafrost sites were Thule, Greenland; Barrow, Alaska; Svalbard, Norway. The discontinuous permafrost sites were Imnavait Creek, AK; Eightmile Lake, Alaska. The data on NEE spanned a time period from 2008 to 2018 for the discontinuous sites and 1998 to 2016 for the continuous sites. The NEE data from each site did not span the entire period but they all fell within the specified range. NEE or CO2 flux data (which is interchangeable) was retrieved from the data sets. If there was not NEE or Carbon Flux measurements, the NEE was calculated by adding ecosystem respiration to the gross primary production (Kramer et al., 2002).
The Thule, Greenland data set and the Barrow, Alaska data set contained CO2 flux that was collected via soil probes and consisted of three different data sets (Konkel, 2017) (Oechel et al., 2001). The date format of each was standardized. The data was also subset for the control treatment type as the original data had experimental treatments that would affect the results of our analysis. The data from Barrow was expunged of any measurements of -9999 as those were readings that were not usable. The Kangerlussuaq, Greenland site measured NEE using a custom-built clear acrylic chamber and an infrared gas analyzer (Post et al., 2016). The data was subset by treatment type in order to avoid any experimental treatments from the study in Post et al., 2016. The Eightmile Lake, AK site measured NEE using an infrared gas analyzer (Schuur, 2018 format was changed to match the Eightmile Lake, AK format and then was added to the Eightmile Lake, AK data.

Data Analysis
The variables selected in this study were permafrost type, month, NEE, and Month x Type. NEE, measured in Micromoles of carbon per Meter squared per second, was plotted against month. The NEE was treated as the dependent variable and the date was treated as the independent variables. Data analysis was conducted in JMP software. Two linear regressions were made, one for the continuous data and one for the discontinuous data, this was to determine the trend in NEE over the last decade and to see if there was a significant relationship between date and a change in NEE. The p-value and R 2 values were recorded for the two tundra types. An ANOVA test was run against the two types of permafrost in order to determine any significance between the changes in NEE in the continuous and discontinuous sites. Figure 2 was created using the location of each site and a blank map using Adobe Photoshop software.

Results
Net ecosystem exchange in tundra is increasing (Figure 1). Both the discontinuous and continuous permafrost zones displayed positive changes in NEE between 1998 and 2016.
Continuous zones exhibited a greater change in NEE than discontinuous zones. The continuous permafrost change is given by the slope: NEE= 0.005896x, with an R 2 value of 0.25 and the discontinuous zones with the slope: NEE= 0.003756x, and an R 2 value of less than 0.0001.

Figure 1. Net Ecosystem Exchange (NEE) of Continuous Permafrost and Discontinuous Permafrost
Zones 1998-2016. There were three sites of continuous and two sites of discontinuous plotted together. The different rates of carbon release and the different rates in change of carbon release between continuous and discontinuous permafrost zones was also found to be statistically significant from the ANOVA test resulting in a P-value less than 0.0001.

Discussion:
Our study revealed important information about the different ways in which permafrost is changing in response to a heating climate. From the analysis that was conducted it is clear that The fact that continuous zones of permafrost are responding more quickly to climate change may indicate that they will begin to resemble discontinuous permafrost zones in the near future and more ice melts and cracks form in the ice and frozen soil. This would be a drastic change as our study found that while continuous zones are changing more rapidly, discontinuous zones are emitting much more carbon into the atmosphere indicated by the larger mean value.
This finding makes sense as there are more areas in the discontinuous permafrost soil and ice to allow trapped carbon to escape into the air (Pollard, 2018). This implies that if continuous permafrost zones are going to begin to resemble discontinuous zones in the future, there will be much more carbon being released into the atmosphere from regions of tundra.
All of this demonstrates that arctic tundra is quickly becoming a source of atmospheric carbon rather than a sink. Despite this insight, there is much more research and analysis that needs to be conducted on arctic permafrost. If we are to more thoroughly understand the implications of continuous permafrost zones responding to climate change more rapidly than discontinuous permafrost zones we would want to collect data spanning further into the past and many more sites that encompass more than just the western hemisphere. We would ideally get data on how both continuous and discontinuous permafrost zones have changed in the last 50 to 100 years in order to see what type of curve the NEE change follows and determine with greater confidence what point the continuous zones will resemble the discontinuous zones. If this study was to be conducted again, care should be taken when compiling data as there were strange outliers in the continuous data sets that had to be expunged due to odd treatment groups being included with the control data.
Overall, our findings are in line with similar studies such as Biskaborn et al., 2019, which demonstrated that arctic permafrost is indeed melting and Schaefer et al., 2011, which demonstrated that CO2 emissions are increasing from permafrost. Our study however built on the knowledge of how carbon emissions are changing differently in the different zones of permafrost. The extent of which this will affect the greenhouse effect was studied in González-Eguino et al., which found that an increase of emissions from permafrost could increase the cost to society by up to 17%. Based on our analysis and understanding, our findings are in line with the conclusion that the additional carbon emitted from permafrost zones will have an effect on the greenhouse effect which will increase as NEE also continues to increase in different permafrost zones. Attention should be placed on the arctic permafrost as a future source of atmospheric CO2 as it could potentially change the course of climate change.
The Connection Between Arctic Permafrost and Global Warming Gap Tundra transition from sink to source but little understanding of how each permafrost type thaws throughout the years 4 Hypothesis Both discontinuous and continuous permafrost zones will show positive net ecosystem exchange (NEE) values over time. In comparison to continuous zones, discontinuous zones will release more carbon due to more plentiful ice cracks that promote thawing; also, discontinuous zones will respond to changes in climate more dramatically due sunlight penetration through ice cracks which warms deeper ice layers.

Conclusion
• Rainfall deposition caused by a major weather event, such as a hurricane, causes an increase in nutrient levels in the lakes affected. • The nutrients affected by the hurricane event could be influenced by the size of the affected lake. • The nutrient levels that will change following a hurricane event increase exponentially and then return to a level slightly above the nutrient level exhibited by the lake in the year before the hurricane event. • Once water nutrient testing can be performed daily more accurate results can be obtained. after Hurricane Jeanne. Prior to the hurricane, the R 2 value was 0.02 and the P-value was 0.0062. After the hurricane, the R 2 was 0.14 and the P-value was less than 0.0001. Figure 2: Nitrogen levels in Lake Okeechobee a year before and after Hurricane Jeanne. Before the hurricane, the R 2 value was 0.11 and the P-value was less than 0.0001. After the hurricane event, the R 2 value was 0.19 and the P-value was less than 0.0001. Figure 3: Nitrogen levels in Lake Istokpoga a year before and after Hurricane Frances. The R 2 value prior to the hurricane was 0.18 while the R 2 value following the hurricane was 0.02. The P value before the hurricane was 0.0037, and the P value following was 0.6552. Figure 4: Nitrogen levels in Lake Istokpoga a year before and after Hurricane Jeanne. Before the hurricane, the R 2 value was 0.21 and the P-value was 0.0013. After the hurricane, the R 2 value was 0.31 and the P-value was less than 0.0001

Results
• A general increase of nutrient levels following hurricane events. • Lake Okeechobee experienced an exponential increase in phosphorus and nitrogen following Hurricane Jeanne. • Analysis of Lake Okeechobee following Hurricane Wilma did not yield a significant change. • Nitrogen levels in Lake Istokpoga were the only nutrient to experience a statistically significant increase following both hurricane events.

Background
• Global climate change is causing an increase in severity of hurricanes.¹ • Productivity of lakes decrease following a major rain event and there's an increase in dissolved organic material. 2 It hasn't been shown how hurricanes affect nutrient levels. • An excess of nutrients can cause eutrophication, decreasing lake productivity, biodiversity and an increase in lake toxicity. 3 Research Question: How does the change in severity of hurricanes affect a lake's nutrient levels? Hypothesis: As the intensity of tropical storms increase, there will also be an increase of nitrogen and phosphorus levels in the lakes following severe hurricanes.