GEOG 302: Biogeography                                                                                                        Spring 2008

---

Assignment II: Future Climate Change, Snowfall, and Forest Response

DUE Friday 8 February, 2008

Reading Background: approximately 30 min.
Parts I-IV (data manipulation and graphs): approximately 1-2 hours (more or less, depending on Excel proficiency). Part V (short answers): less than 1 hour.

Goals
Background and Purpose
Part I: Downloading the data and creating Excel files
Part II: Calculate state-wide averages
Part III: Compile state-wide averages from each scenario and calculate proportional changes in snow amount
Part IV: Plotting trends in snow amount
Part V: Interpreting patterns and predicting forest response to alternative emissions scenarios

---

Goals

1. Learn about available data from the International Governmental Panel on Climate Change (IPCC) and alternative scenarios for future climate during the 21st century.

2. Download data in their original format, manipulate data in a spreadsheet, and generate graphs.

3. Analyze graphs of future annual snow amount in Montana and Washington, and interpret similarities and differences between scenarios and between states.

4. Make inferences about the impacts of climate change, and specifically snow fall, on the future distribution of several conifer species studied by McKenzie et al. (2003) in Washington State.

---

Background and Purpose

The average annual temperature in the United States increased by almost 0.6 degrees Celsius (1 degree Fahrenheit) over the 20th century and precipitation increased by 5-10% nationally, largely the result of heavier downpours. Scientists use computer models to describe our current weather patterns and to predict how our climate might change in the future. Two of the most widely used models each predict that the United States will become warmer over the next century, by roughly 3 to 5 degrees Celsius (5 to 9 degrees Fahrenheit). These estimates exceed the projected global increase. 

Across the Western US, snow fall during the winter is a critical resources, and increased temperatures in the future will have important impacts total snow amounts. Snow represents a natural reservoir   which releases water during the late spring and early summer. Winter snow fall thus strongly impacts summer water supplies, with implications for irrigation, drinking water, fish and wildlife habitat, vegetation, and forest fire activity. 

In this exercise, you will access projected climatological data for Montana and Washington developed for the IPCC's Fourth Assessment, released in 2007. The authors of this report shared the 2007 Nobel Peace Prize with former Vice President Al Gore. If you are unfamiliar with this report, please read the Summary to Policy Makers. In Part I-IV of this exercise you will import predicted snow amount data into a spreadsheet application, manipulate the data, and create two graphics illustrating predicted snow amounts for Montana and Washington states under three alternative emissions scenarios. In Part V you will analyze the datasets to interpret variability between alternative future scenarios and between states. Finally, based on the importance of snow fall for the presence and absence of certain conifer species (from McKenzie et al. 2003), you will make predictions about the potential impacts of future changes in snow amount for conifer species and forest communities.

Our Dataset and Emissions Scenarios  Our dataset was downloaded from EOS-WEBSTER (see acknowledgments), which provides the following description:

"This collection contains global historical climate and future climate projections produced by the National Center for Atmospheric Research (NCAR) for the Intergovernmental Panel on Climate Change (IPCC) 4th Assessment Report. There are 4 future climate scenarios based on different emissions of CO2 and other gases as well as socio-economic factors. Estimates were generated by the NCAR Community Climate System Model (CCSM3). Scenarios include SRESA2, SRESB1, SRESA1B and Commit. Data are annual and decadal values at a spatial resolution of 1.40625 degrees. Date range: 1870-01-01 - 2099-12-31"

Table 1. Description of the datasets used in this exercise.  For more information on IPCC Emissions Scenarios, see: IPCC Special Report on Emissions Scenarios

Dataset	Description	IPCC Name	Dates	Download
Climate of the 20th Century	Atmospheric CO2 concentrations and other input data are based on historical records or estimates beginning around the time of the Industrial Revolution.	20C3M	1870 – 1999	Montana_snow Washington_snow
Year 2000 CO2 maximum (Commit)	Atmospheric CO2 concentrations are held at year 2000 levels. This experiment is based on conditions that already exist (e.g., “committed” climate change). Details	Commit	2000 – 2100	Montana_snow Washington_snow
550 ppm CO2 maximum (SRESB1)	Atmospheric CO2 concentrations reach 550 ppm in the year 2100 in a world characterized by low population growth, high GDP growth, low energy use, high land-use changes, low resource availability and medium introduction of new and efficient technologies.	SRESB1	2000 – 2100	Montana_snow Washington_snow
850 ppm CO2 maximum (SRESA2)	Atmospheric CO2 concentrations reach 850 ppm in the year 2100 in a world characterized by high population growth, medium GDP growth, high energy use, medium/high land-use changes, low resource availability and slow introduction of new and efficient technologies.	SRESA2	2000 – 2100	Montana_snow Washington_snow

You will find a variety of other datasets on predicted  temperature, precipitation, vegetation,  etc., at the EOS-WEBSTER site.

A Cautionary Note About Model Projections “Scientists use computer models to describe mathematically the various factors that affect our climate and to characterize the interrelationships among these factors. These computer simulations can also be used to make projections as to how climate may change in the future. The data sets used in this exercise are derived from a model developed by the [National Center for Atmospheric Research’s (NCAR) Community Climate System Model, version 3 (CCSM3)]. Models are simplified representations of our physical environment. As such, there are limitations and uncertainties within all computer models. Some of these uncertainties are due to our incomplete understanding of the key processes that affect climate or the difficulty of incorporating these processes into the model. As our understanding of these processes improves and we are better able to describe these processes mathematically, these kinds of model uncertainties should diminish. The limitations and uncertainties inherent in scientific models do not make these models useless or invalid. One test of the effectiveness of computer simulations is how well they predict historical climate shifts since we know how climate changed in the past. Many models have successfully reproduced these historical climate changes. Models are also an essential tool for improving our understanding of Earth system science. We simply have no better means for predicting how climate may change in the future. Still, it's important to keep the inherent uncertainties and limitations of computer modeling in mind when using the model output data in this exercise. Projections of temperature and precipitation changes over the next century should be regarded as possible scenarios of what might happen and not an absolute assertion of what is to come. (For more information on climate models see: Climate Model Scenarios and Uncertainties).”

Excerpt from the Earth Exploration Toolbox, where this exercise  was modified from: http://serc.carleton.edu/eet/climate/index.html

Emissions Scenarios Used by IPCC


Figure 1. Atmospheric CO2 concentration scenarios used to drive the different IPCC scenarios. "MIN" pattern of CO2 emissions is used for the "Commit" scenario (see below).

Acknowledgments

The following exercise utilizes data obtained from the EOS-WEBSTER digital library of Earth Science data (http://eos-webster.sr.unh.edu) and is modified from a case study presented in the Earth Exploration Toolbox (http://serc.carleton.edu/eet/climate/index.html)

---

Part I: Downloading the data and creating Excel files

1. Create a folder to house the datasets used in this exercise: Create a folder anywhere you like on your computer. The datasets will be saved in this folder.

2. Download the data: Right-click on the link to each dataset (below) and select "save target as" (Internet Explorer) or "save link as" (Firefox). Save each file as a text file (suffix of ".txt"). Once done, you should have eight files saved in the folder you created in step 1.


3. Import the data into Excel: Import each of the eight files into one of eight separate Excel spreadsheets, following the directions below. You should have one spreadsheet for each datafile.  

1. Open Excel, and from within Excel, open one of the eight data files using the Open command, under the File menu. All of the data files are comma delimited text files, so make sure you change "Files of type" to read "All Files (*.*)". 

    2. Our files are delimited (by commas), so be sure "Delimited" is selected and then click "Next".


    3.    Select "Comma" as the delimiter (and unselect "Tab"), and then click "Finish".


    4. Your file should now have individual values in all cells below row 10. It should look something like this:



    5.    Finally, save your file as an Excel file (.xls) by selecting Save As under the File menu.Your file type should be "Microsoft Office Excel Workbook (*.xls)" (not necessarily what's shown below). 

---

Part II: Calculate state-wide averages

Each data file contains information for the entire state of Montana or Washington, at 1.4 degree grid cells (this defines the spatial resolution of this dataset). To make state-specific comparisons, we need to average the predicted values from the grid cell data.

1. Place your cursor in the blank cell in the first column of your worksheet below the MT or WA data and click on this cell. As you will be creating an average in this row, label this row with the name of the Emissions Scenario, and the state: e.g. MT A2 scenario,  MT commit scenario, MT B1 scenario, or MT historical.

2.    Place your cursor in the cell below the column of grid cell data you want to average. In the example below, this column corresponds to the year 1870. In all cases, it will be the third column (as the first two correspond to longitude and latitude). Click this cell. To calculate an average from all the grid cell points, use the Excel function "average", which does just what it suggests (calculates an average of the cells you refer to). Type into this blank cell the text: =average(


3.    The equals sign before the word tells Excel that we want to use a function, and in this case the function "average". Now select the column that corresponds to the first column of data. Be careful to only select the snow amount data and not the row containing the Year label. Once you select the column, close the formula by using a close parenthesis ")". Your formula should look something like the one below (but the range will vary).



4.     Once you have the formula for the first column, you can copy and paste the formula into all the cells to the right, corresponding to each year in the dataset. Place your cursor on the cell you just put the average formula into, then select Copy from the Edit menu. Scroll and select the adjacent cells (by holding down the left mouse button) in this row that correspond to the reaming years in the dataset, then choose the Paste command under the Edit menu.



5.    You should now have a value for each year representing the state-wide average predicted snow amount. Your file should look something like this:



6.    Save your work, and repeat these steps for the seven remaining datasets. Hint: you can copy and paste the formulas you used to calculate state-wide averages between Excel files. This can save you some time. In this case, you want to copy the formula, so don't use Paste Special.

---

Part III: Compile state-wide averages from each scenario and calculate proportional changes in snow amount

We need to compile the forecasted (i.e. A2, B1, and commit Emissions Scenarios) snow amount data from each state into one file, so we can create  graphs containing each time series.

1. Download the files MT_snow_data.xls and WA_snow_data.xls (by clicking on file name). These files contain space for the state-wide averages you calculated in Part II (rows 3-5). Additionally, there is room to put snow amount ratios, which we will calculate based on the average snow amount from 1980-1999 (below).

2. For each state, and for each emissions scenario, highlight the state-wide average values for the years 2000-2099 and select Copy from the Edit menu. Move to the proper row in the file you downloaded in step 1 and select Paste Special from the Edit menu (do not simply paste into the new file, as this will paste the formulas, not the values). You must convert the formulas you made in Part II into values by choosing "Values" from within the Paste Special window. 



3. After you complete step 2 for each scenario, your file should look like this:



4. We want to calculate relative changes in snow amount so we can (1) compare changes between Montana and Washington, and (2) draw inferences into the impacts of climate change on the conifer species studied by McKenzie et al. (2003). Our snow amount ratio will relate the annual amount of snow in the future to the average annual amount of snow between 1980 and 1999. This 20-yr period serves as our "base period": changes in the future are all relative to this period. For example, a ratio of 1 represents 100% of snow amount during the base period, whereas a ratio of 1.5 represents 150% and a ratio of 0.5 represents 50%. Open the Excel file that contains the data from the 20th century scenario (20C3M),which you created in Part II. Use the Excel function "average" again to calculate the average (state-wide average) snow amount for the years 1980-1999.

5. For each state, copy the resulting average value and return to the file we left in step 3 (e.g MT_snow_data.xls). Use the Paste Special command under the Edit menu again, this time pasting the value into the cell C12, labeled for the 1980-1999 average.

6. To fill in the values for the snow amount ratio, we need to construct one more formula in excel. Select cell C9 (in the example below) and enter the following text: =C3 / $C$12
This formula tells Excel to divide the value in cell C3 (MT_snw_commit data for the year 2000) by the value in cell C12 (the average snow amount from 1980-1999). The "$" symbols tell Excel to always refer to this cell, no matter where the formula is copied. Without the $ symbol, Excel will look "6 cells above" the cell where this formula is pasted. With our $ symbols, we can copy the value in cell C9 and paste it into cells C10 and C11, and then we can copy the three rows of data for the year 2000 and paste them into the same rows corresponding to the years 2001-2099. Note: you can copy, then highlight multiple cells, and then paste. You should not have to paste the formula into individual cells (this would take way too long!). 



7. Save your data files. You now have all the data required to construct our plots, and they are organized in two files, one for each state.

---

Part IV: Plotting trends in snow amount

At this point you should have two files (MT_snow_data.xls and WA_snow_data.xls) that contain predicted snow amounts for each of the three emissions scenarios, in both mm / yr and as a snow amount ratio. In Part IV you will plot the three time series for each state on a single graph (for a total of two graphs, one for each state).

1.    Within your data file, select the data range intended for plotting, as indicated below - you want to highlight the columns corresponding to years through 2099. Next, select the Chart Wizard icon (or alternatively select Chart from the Insert menu). This will bring up the Char Wizard window (shown below). Select the Scatterplot chart type and the chart sub-type 'Scatter with data points connected by smoothed lines without markers". Then click Next.



2. You should then see a preview of your graph, which should look something like this below. Click Next.



3.  On the next screen, be sure to add labels to your x axis and y axis and give your graph a descriptive title. ["CE" refers to "common era" and is often used in scientific medium in pace of "AD" (after death) given its less religious connotation]

4. Place your chart in the current sheet or make it its own sheet, as you choose. If you place it in the sheet, be sure to modify the size of the chart to have an aspect ratio that makes sense. You want to be able to easily observe the differences between the three alternative scenarios. Also, make sure the size of your MT and WA charts are the same, so you can easily make comparisons between them. Click Next.



5. After your graph is created, click on the individual lines (series) to change their colors to something more useful and eye-pleasing than the Excel default: make the commit scenario green (the best option going), the A2 scenario red (danger!), and the B1 scenario blue (middle of the road).



6. Make the figure background all white by turning off the Area color in the Format Plot Area window.



7. Make sure the graph makes sense: check the significant digits on the y axis (there should be 1-2 decimal points, not 4). You can change these options in the "Numbers" tab, in the Format Axis window. Access the Format Axis window by double clicking on the number on the axis you are interested in. Also select the Scale tab and change the x-axis scale to go from 2000 to 2100. NOTE: Be sure to have the same scales for both your MT and WA graphs - this facilitates easy comparisons between the two states.



8. Finally, copy each graph (select Copy, under the Edit menu) and paste it into a text document (e.g. MS Word). Annotate your figure with a figure legend, which should go *below* the figure. Make the Montana data Figure 1 and the Washington data Figure 2. 

---

Part V: Interpreting patterns and predicting forest response to alternative emissions scenarios

You now have predicted snow amounts for both Montana and Washington state under three alternative emissions scenarios considered by the IPCC. In this section you will (1) describe the patterns in these datasets, and (2) use the predicted snow amounts for Washington State to make some prediction about he potential future for mountain hemlock (Tsuga mertensiana), lodgepole pine (Pinus contorta) and whitebark pine (Pinus albicaulis). Using your graphs and the figure below from McKenzie et al. (2003) paper, answer the questions below.




Figure 2. Data from McKenzie et al. on relationship between winter precipitation and the probability of occurrence for three conifer species.


Questions:

Answer the following questions in the same text file containing your final figures. Each answer should be no longer than 150 words. Be clear and concise, and justify your answers.

1. What does the y-axis on each figure represent? What does a value of 1, 2, and 0.5 mean? (1 pt)

2. For Montana, how do the predicted snow amounts differ under the three emissions scenarios?  Focus on three temporal scales: total amount of snow over the 21st century, decadal-scale trends (i.e. trends over decades), and the inter-annual (i.e. between-year) variability in the different time series. (2 pt)

3. Are these differences present throughout the 21st century, or do they become more apparent after a particular date?  After looking at the patterns of CO2 emissions for the three different scenarios used in this exercise (i.e. Figure 1, in Part I - Background), explain why this pattern (or absence of a pattern) might exist. (2 pt)

4. What is the main difference you see when comparing the Montana predictions to the Washington predictions? Given the location of these two states (and their temperature differences - hint: WA has warmer winters than MT), what processes do you think accounts for these differences?  (2 pt)

5. Based on the predictions for Washington State and the assumption that these predictions apply to the region studied by McKenzie et al. (2003), what would you predict about the future presence of mountain hemlock under the three different scenarios? How about lodgepole pine and whitebark pine?

Note: the average amount of winter precipitation received in the Grizzly Bear forest studied by McKenzie et al. is currently 65 cm. For this question, assume that all the winter precipitation in the Grizzly Bear forest falls as snow (and therefore would be subject to the changes predicted by your snow amount figures).  This is a reasonable assumption, given the location of this study.  (2 pt)

6. On the eastern slopes of the Cascade Mountains (in WA), mountain hemlock and whitebark pine are found within the same "community". Based on predicted snow amounts, what do you predict will happen to this community under the A2 scenario? What does this suggest about the future of modern communities under some climate change scenarios? (2 pt) 

7. Snowfall in Washington State is predicted to decrease in the future primarily because of increased temperatures, rather than a decrease in winter precipitation. From the perspective of a tree, what is the difference between receiving rain vs. snow in the winter? How would this impact growth during the following growing season? (2 pt)

8. List and explain 5 potential limitations to the inference you described in #5 and #6, considering both the CCSM3 climate model used and the niche model presented by McKenzie el al. (2003). Consider aspects of spatial scale, temporal scale, and species interactions (2 pt)

---