I have attached the assignment as well as the csv file that is helpful to write the code.
COMP1730/6730 S1 2020 — Project Assignment COMP1730/6730 S1 2020 — Project Assignment Matthew Alger 06-04-2020 Important • The assignment is due 11:55 PM Sunday in week 11, May 31. • This assignment needs to be completed by yourself (i.e, not in a group). You can share ideas (with proper attribution) but your code and report must be your own. • Include your university ID in every file you submit. • This assignment is worth 15% of your grade for COMP1730/COMP6730. Overview In this assignment you will be doing a short data analysis and modelling task using some real-world geographical data. It is different to the homework assignments you have done up until this point in that for almost all of the questions, there is no single “right” answer. You are also not given any tests against which to check your answers (although you are encouraged to write your own to help you test the correctness of your functions). Because there is no single “right” answer, it will be important to justify the decisions and choices that you make while completing the assignment. This is important since it allows anyone relying on your results and conclusions to understand how they were obtained and whether they are suitable for a particular purpose. Lake George, or Weereewa (in Ngunnawal), is an intermittent lake to the north-east of Canberra. It is well-known for having dramatic changes in water level. When full, it’s around 145 km2—20 times bigger than Lake Burley Griffin! But most of the time in contemporary history, it’s been fairly empty. This unusual behaviour has given Lake George quite the reputation and even some persistent myths and legends, including everything from a secret underground link between Lake George and Mount Gambier, to bunyip and alien sightings. Of course, these legends are (probably) not true, and the fluctuating water levels of Lake George are explained by its unusually shallow geology, evaporation, and rainfall. In this assignment you will investigate the water level of Lake George and develop two models of how it fills and ebbs over time. The Data We have provided you with a CSV file assignment_lake_george_data.csv containing meteorological data for the Lake George region from 1990–2018, as well as the area and volume of the lake over time. This is made from data from the Bureau of Meteorology, Digital Earth Australia, and the CSIRO. You may recall CSV files from Labs 6 and 8. They are text files made up of rows of data, with each column separated by commas. There are 9 columns in this dataset: • Date, in YYYYMM format; • Volume of Lake George, in litres (as at the end of the month); • Area of Lake George, in square metres (as at the end of the month); • Total monthly solar exposure, in megajoules per square metre; • Total monthly rainfall, in mm; 1 Figure 1: Geology of Lake George (Jacobson & Schuett, 1979). 2 • Average daily maximum temperature, in degrees Celsius; • Average daily minimum temperature, in degrees Celsius; • Average daily relative humidity, as a percentage; and • Average monthly wind speed, in metres/second. The file looks like this (but with more rows): date,volume,area,solar_exposure,rainfall,max_temperature,min_temperature,humidity,wind_speed 199001,190850270907.59,141952429.50,22.10,22.80,22.82,8.44,37,1.64 199002,184683568988.20,141499648.55,19.40,83.30,22.32,8.41,40,1.26 199003,180741119308.03,141298750.00,16.30,17.60,21.86,8.38,42,1.39 This would correspond to the following table: Date Volume Area Solar exp. Rainfall Max temp. Min temp. Humidity Wind 199001 190850270907.59 141952429.50 22.10 22.80 22.82 8.44 37 1.64 199002 184683568988.20 141499648.55 19.40 83.30 22.32 8.41 40 1.26 199003 180741119308.03 141298750.00 16.30 17.60 21.86 8.38 42 1.39 The Task You are provided with assignment_template.py, which contains the basic functions of the assignment. The functions are incomplete. In this assignment, you will fill in the blanks and complete the missing functions. You will also write a short report about your functions and decisions. Question 1: Reading the Data Write a function that takes the file path of the dataset as input, reads the data, and returns the dataset in a suitable format. The assignment template contains a function for you to fill in: def read_dataset(filepath): pass pass means “do nothing”, and you should remove it when you fill in this function. To load the data, you can then run data = read_dataset('assignment_lake_george_data.csv') as long as the CSV file is in the same directory as your assignment file. You should read the data from filepath, and return it in an easy-to-use format. This can be any data type or data structure that you like, as long as it makes sense for the tasks you will be doing later in this assignment. You will be using this returned value in all other questions of the assignment, so make sure your choices here support your later solutions! Hint - have a look at the remaining questions before deciding on what format to load your data in! Question 2: Statistics about Lake George Write four functions that output (return) the answers to the following questions: a. What is the largest area covered by the lake? b. What is the average volume of the lake? c. What month and year had a rainfall closest to average? d. Which month is the hottest on average? The corresponding four functions are outlined in the assignment template: def largest_area(data): pass 3 def average_volume(data): pass def most_average_rainfall(data): pass def hottest_month(data): pass All of these functions should take as an argument the Lake George data in the same format you loaded it in Question 1 (for example, if you loaded the data as a list, then data is a list here). Question 3: Lake George Topography Lake George is notable for being very shallow. We want to identify points where the area increases quickly with little volume added, and vice-versa. We have written a function area_vs_volume which should plot the area of Lake George against its volume. Our function doesn’t work though, and even if it did, the plot is so poor it doesn’t really help us answer our question. Fix the function so that it works on your data, and improve the plot so it makes it easy for someone looking at it to identify the points we are interested in. The function to fix is in the template: def area_vs_volume(data): areas = [] volumes = [] plt.plot(areas) plt.plot(volumes) plt.show() You will be marked on how good the plot looks (and how appropriate it is for the task) as well as whether the code works. You are free to pick a completely different plot type if you think it is more appropriate. You should include a copy of your plot in your report. Hint - have a look at some of the visualisation tips included in Lecture 12. Question 4: Modelling Lake George Lake George is an interesting lake because it is endorheic, meaning that it has no outflows: water only leaves the lake through evaporation. While the physical process of evaporation can be very complicated, in this question we will assume a much simpler situation. We will also assume that Lake George has only a single inflow, i.e. rainfall (thus the secret underground passage to Mount Gambier is outside the scope of the assignment). We want to develop a model for how Lake George fills and ebbs over time. A model is a programmatic description or simulation of a physical process, like how much water is in the lake after it receives some amount of rain. In this question, you will build two different models of Lake George: a simple one which assumes that evaporation is constant, and a more complex one which assumes evaporation is dependent on some other variables. Evaporation and rainfall are both measured in millimetres (mm)! 1 mm of rainfall corresponds to 1 litre of rain falling per square metre of area. Similarly, 1 mm of evaporation corresponds to 1 litre of water evaporating per square metre of area. 4 Part (a) Use the monthly rainfall data to estimate the volume of Lake George for each month in the data, assuming a constant evaporation rate each month. Write a function lake_george_simple_model which takes the Lake George data in your format, as well as the assumed rate of evaporation (in mm/month), and returns a list of modelled lake volumes (in litres). We have outlined this function in the template: def lake_george_simple_model(data, evaporation_rate): pass Hint - in order to convert both water gained from rainfall (in mm/month) and water lost to evaporation (also in mm/month) into volume, you will need to make assumptions about both the catchment area and surface area of Lake George. You can always start by assuming they are both equal to the maximum surface area of the lake (which you calculated in Question 2) and then refining them from there if you want. Another hint - you will need a starting point for your calculations. You can use the first record in the data set as the starting point for your models. We have also given you a function plot_volumes to help visualise the results. You can use it like this: plot_volumes(lake_george_simple_model(data, evaporation_rate)) Try plotting your model for different values of the evaporation rate. For example, if you set the evaporation rate to 55 mm/month and use plot_volumes, you should see something like this: Figure 2: A simple model for Lake George with 55 mm/month evaporation. 5 Part (b) While our simple model will give some very rough answers, it might be too simple to be of much use for real analysis tasks. For this question, instead of assuming evaporation is constant, use the following equation to calculate the evaporation rate E (in mm/month): E = −3Tmin + 1.6Tmax − 2.5W + 4.5S − 0.4H where T is the temperature (in Celsius), S is the solar exposure (in MJ/month/m2), W is the wind speed (in m/s), and H is the humidity (as a percentage, i.e. as a number between 0 and 100). Write a function lake_george_complex_model that uses this evaporation rate instead of assuming the rate is constant. We have outlined this function in the template: def lake_george_complex_model(data): pass Question 5: Evaluating a Model We can make as many models as we like, but if there’s no way to quantify how good those models are, then we have no way to choose which model we should use. Write a function evaluate_model that takes in your Lake George data and a list of volumes (in litres) and returns a float that indicates how bad your model is at estimating the real values. This float is called the model error and it should be equal to zero if and only if the model volumes are all exactly the same as the real volumes. We have provided an outline in the assignment template: def evaluate_model(data, volumes):