Guidelines for writing the major report
GEOS2821, 2025
These are some guidelines to aid you in writing your major report. These focus on style and structure. This is not the only way of structuring it so do not let it stop you from using a better approach (but please pay heed to the term “better”).
There is a lot here so be sure to read it several times, with gaps between to allow your subconscious mind to cogitate on its contents.
Marking criteria and grades
General marking criteria have been given to you in the course outline (additional details) document, as have grade descriptions following the SOLO taxonomy. Further to that:
The basis for a pass mark is to show that you have followed the instructions and that you understand what you have done.
More marks will be awarded for people who go beyond the instructions and show a greater understanding, including integration of material and concepts.
A very well written report that shows a high quality understanding of the concepts, but with relatively simple analyses, will receive a maximum grade of distinction. A report presenting excellent analyses that improve substantially on the instructions, but containing issues with the presentation and writing, will also receive a maximum mark of distinction.
Achievement of a high distinction requires both well executed analyses and a high quality report.
Word count
The word count should be 4500-6500 words, including the executive summary but excluding the bibliography. This supersedes any figures given in the course outline. The upper limit is to allow enough space for people who have implemented many innovations. The lower limit is a guide to ensure people provide enough details.
This works out to 1500-2000 words each for a three person group so is not particularly onerous.
Minimum number of references
You are to have at least ten peer-reviewed geospatial references, i.e. these must be directly related to GIS and/or remote sensing. This is in addition to other general references such as sources of model parameters. As a rule-of-thumb you should have about 20 references in total. Further details are below.
Key points from recent years
Many of these are reiterated in the next section. This is an indication of their importance.
• This is a GIS and RS report first and foremost. That must be your focus.
• Use figures to explain concepts.
a. You can re-use figures from textbooks and articles. Just remember to cite the course.
b. Use ArcGIS Pro to make proper maps. Do not use screenshots.
c. Build your own location diagram using ArcGIS. Remember that this must include a false colour composite of one of the Sentinel data sets you used early in the course. You can adjust the stretch and use infrared bands to show more variation. Its resolution should be sufficient for maps of the study region extent.
• The consultancy brief clearly states that the methods need to be described in such a way that someone who does not have access to ArcGIS can replicate what you have done using a different software package.
• Define GIS and explain why you are using it. The same applies to RS.
• Use a logical structure so ideas build on each other. This does not mean everything is sequential but, if you use fuzzy logic across several models, then clearly it needs to be explained before all of them. The same applies to data sets used across the models. This is why the guidelines note this in the structure.
• Cross-check your final site locations with other available imagery, for example basemaps in ArcGIS Pro or images in Google Maps. We know the vegetation data are not 100% accurate so use other data to help constrain your conclusions. People have in the past recommending swamps as good sites to build, which is clearly wrong.
• Explain fuzzy logic and why we are using it. Use figures.
• Explain map algebra. Use figures.
• Explain the fuel load data sets. They are derived from field sampling on the field trip. And you can refer to the field sampling as being done by “field crews of sub-contractors” instead of “students” (one way of presenting the work as a professional report).
• Think about your cell resolutions and what is the minimum resolvable feature.
• Slope in these models always uses the DEM. It is topographic slope so is a function of elevation.
• Think about the accuracy of your data sets and how these can be used to help define some of your distance parameters.
Style
(These are numbered to allow cross-referencing).
1. Write this report as if you will later use it as part of a job application to demonstrate your skills. What would you want a prospective employer to see?
2. Describe your methods in sufficient detail that someone else can repeat what you have done but using a different operating system and GIS software. Just saying you combined data sets is not sufficient. Give your parameters and describe the algorithms. Do not merely list the tools you used. Be sure to state that you used ArcGIS Pro 3.x, though (e.g. 3.3, as appropriate). The same applies to ENVI where you use your remote sensing products from the first part of the course.
3. Focus on the GIS and remote sensing side of things, and refer to the relevant background theory. That said, you do not need to provide the background theory behind everything you do. For example, for the fire model you need only state that you used the method of Noble et al. (1986) and then state the parameters you used. You do not need to summarise the paper. If there are relevant criticisms you are aware of then state them, citing any relevant literature and how you have adapted, or could adapt, your model for their potential effects.
4. This work is framed as a consultancy report so the presentation style. is slightly different from a scientific report. However, you still need to convey your ideas effectively and refer to the relevant literature to support your statements.
5. Please note that, while the word limit is a maximum of 6500 words including the executive summary, this is the MAXIMUM. Needlessly long reports indicate a lack of time taken to draft your report such that it clearly conveys your ideas.
6. This is a piece of formal writing. Avoid conversational terms and phrasing.
7. Be direct and concise. This sounds simple but actually requires some effort on your part in drafting the report. A basic principle is that it is easier to correct than to create. Accept that your first draft is likely to be of low quality but realise that you can only improve it once you have written it. This means that many of the ideas in the first draft will be poorly formed or conveyed, possibly even incorrect. However, once you have them on the page and out of your brain you can begin correcting or improving them. You will also have removed the bad ideas from your system, as writing something down (and sometimes reading it aloud) forces you to think it through more thoroughly without blithely skipping over any gaping chasms of flawed logic.
8. Ask a non-expert reader to read your report before you submit it. Ask them to flag anything they have to read more than once to understand, or where the writing is awkward. This is the “intelligent person test”, where an intelligent person who is not an expert in the field will still be able to understand the intent of your writing.
9. Explain fuzzy logic and why you are using it. This should normally be in a section before you describe the individual models, as it is used across several of them. This is important, hence it is also mentioned above in the feedback from recent years.
10. Describe model components where necessary and give formulae for the models where appropriate. For example, the detailed derivations of the parameters used in the fire model can be located in an appendix if they are needed, but the main calculations need to be present in the main document.
11. Do not list the limitations as a separate section at the end. This makes it look like it is an afterthought, which it should not be. Note any limitations in the document as they become relevant. In most cases this will be as you describe each model component, NOT as a subsection at the end of each model section.
12. Be specific when referring to quantities. Terms like “a few”, “a lot” and “a couple” are vague and loose terms. If the distance is 10 m, then say 10 m. If you are not sure of the exact value, or there is associated uncertainty, then say “approximately 10 m”. “There are more than 100 ocelots in the study area”, not “There are lots of ocelots in the study area” .
13. Do not be over-precise with your numbers. For example, fire model results should be to the nearest 100 kW/m. Your candidate areas for the extension should be in increments of 100 m2, as the cell is the minimum unit. If you have higher precision than this then you have probably simplified your polygons when converting the raster, which you should not do. This might look attractive but it is a distortion of your results.
14. Use the active voice in preference to the passive voice. Overuse of the passive voice can act as a soporific, and a sleepy reader is not what you want. A simple mnemonic is that “The cat sat on the mat”. Nobody ever says that “The mat was sat on by the cat”.
15. Avoid abbreviations in formal writing. For example, use “do not” instead of “don’t”, “cannot” instead of “can’t”.
16. The word “it” is an exception to the rule when using possessive apostrophes. “its” is the possessive form, while “it’s” is an abbreviation for “it is”.
17. Do not confuse the words “complement” and “compliment”. The former provides balance to something, such as a fuzzy membership plus its complement will equal 1. A compliment is when you tell someone that their hair looks nice.
18. Topology and topography have very different meanings.
19. Avoid singular/plural disagreements: “criteria” and “phenomena” are plurals of “criterion” and “phenomenon”, respectively.
20. Use Australian spelling. This means “neighbours”, not “neighbors”, and metres not meters for distances. (Note that “perimeter” is the correct spelling – it is not a distance unit, similar for “meter” when describing something a measuring device such as a water meter).
Cross-referencing
21. A good strategy when writing is to minimise any forwards cross-referencing (where you refer to a section later in the report). If your argument depends on information from a section later in the report, then you should perhaps have already said it. You cannot put a roof on a house until you have built the frame.
22. Make sure you use backwards cross-referencing to direct the reader to where you have already covered a relevant point. Phrase this in the same way as for figures and tables.
23. When referring to figures and tables, do not use phrasing like: “Figure 1 shows where all the ocelots are in the study area, including their names and identifying marks. They are widely dispersed. The names and identifying marks were identified by Aardvark and Genyornis in 2015.” This is awkward and clunky. A better approach is to make your statement and then refer the reader to the relevant figures, tables and references. For example: “The ocelots are widely dispersed across the study area (Figure 1). Identifying marks and names for all ocelots have been described by Aardvark and Genyornis (2015) and are listed in Table 1.” Now the reader knows the ocelots are widely dispersed, and knows to look at Figure 1 to see how much this is the case. They also know to look at Table 1 for names and identifying marks, hopefully linked to the map through appropriate symbols (this should be noted in the caption). Finally, they can read the publication by Aardvark and Genyornis (2015) for more details beyond what you have said or used. They can expect to see its full publication details in the reference list at the end of the document, NOT in footnotes when using the author/year referencing system. You do not normally need to prefix references with terms like “refer to” or “see” .
24. The captions for the tables and figures should allow the reader to quickly assess what the figure is about and should not contain any information that is not also in the main text. They are essentially a short summary of what you want the reader to know once they have looked at the figure or table.
Maps
25. Plot the identified sites using open polygons if you want to simultaneously show the value of some other variable such as solar radiation or slope. If your polygons are plotted using a solid fill then the reader cannot see what you are trying to show, and will perhaps wonder what you are trying to hide.
26. Add feature data to your maps so the reader can more easily cross-check between them. Roads and water bodies for this report. A reader who is unfamiliar with the study area be able to understand very little from these maps:
References
27. The minimum number of peer-reviewed GIS and remote sensing related references is 10. You should have approximately 20 references or more once you have also cited literature on fire, conservation and the like.
28. Only peer reviewed literature will be considered towards the above number. This is a trivial task given the many references provided with the lab and lecture notes. Note that you should still list any non-peer reviewed sources where used.
29. Do not cite lecture and lab notes. The same applies to Wikipedia and similar web sites.
30. Do not cite the software documentation when describing concepts. Read and use the references cited therein.
31. You should refer to web sites you have gleaned data from. However, these will not count towards the minimum reference count.
32. Do not list any reference you have not cited.
33. List all references you have cited.
34. Spell references correctly. For example it is Noble, not Nobel for the fire model.
35. In-text citations should use an Author/Year style. For example “blah blah some important point (Smith and Jones, 1987)” or “…as noted by Smith and Jones (1987)”. See also the course outline.
36. Format your references consistently. Choose a journal style and follow it. It is easiest to choose a simple one that does not italicise journal names and the like, as that reduces the risk of errors. If you use a software tool like EndNote or Mendelay for your references then it will take care of formatting for you (as long as the database is correct).
Report Structure
Spend some time planning your structure before you start writing. This will make it easier to edit your first draft, saving time in the correction stage (see point7).
You are likely to have a series of main sections, with a subsection for each of the models. In this case, you would have something like the following structure. Note that the text in each section represents suggestions for what you need to cover. Sentences in inverted commas are phrasing suggestions.
Executive summary
State clearly what you were required to do, what you did, and what your recommendations are. This is to be no more than half a page, 12 point font (approximately 250-300 words). You should refer to an accompanying map with your recommended locations. This should be on the same or the next page.
Keep in mind than the executive summary is written for someone who has about 30 seconds to read it before they decide whether or not to pass it on to a subordinate for a more detailed reading. A rule of thumb is that a good executive summary will cover all the important details while being read aloud on a single breath.
Introduction
State the aim of the work and the reason for doing it.
The consultancy brief is the statement of what to do. The analyses and report are where the work and results are described.
Give the location of the study area using a location diagram. Do not assume your reader knows where Smiths Lake is, let alone Sydney.
Say why you are using GIS and Remote Sensing to address the brief.
There is no need to list three criteria then list three models. In the case of this report there is a one-to-one match, so one need only say that each criterion is addressed using its own model.
Key data sets and software
Be clear that your analyses were conducted using ArcGIS Pro version 3.x and ENVI 6.x (adding further details if you used other software). (See also point 1 above).
This text is concise (but do not copy this text as it would constitute plagiarism, and also be incorrect since you did not use Landsat or generate a classification for use in the report): “The landcover classification was generated from a Landsat 8 image taken on XX-Jan-2023 using an unsupervised classification method (ISODATA, ). Further details have been given as part of phase 1 of this consultancy.”
State what cell size you used. And how do the cell sizes of the DEM and the Sentinel images compare? What are the possible issues of any resampling?
It is important to note that the MGA56 coordinate system was used, but equally one must justify why it is used.
Analyses
Explain fuzzy logic and why you are using it. Remember that fuzzy logic softens the impact of error (but see also chapter 13 of Burrough et al. 2015).
Do not forget to explain the process of fuzzy overlay. Consider the unfortunate reader who has little familiarity with your methods and who then has to wonder what a fuzzy- OR operation is.
Pay close attention to the values in your layers. Some people in previous years did not pay attention to their fuzzy values and how they were combined, so ended up recommending sites on steep slopes that were far from the roads.
For this work, there is no reason to use anything but fuzzy-AND, fuzzy-OR and fuzzy- Complement in the overlay process.
Use a different subsection for each of your three main models, including any variations.
Use an additional subsection for the combination of these models. You might find that several sub-subsections are necessary for each, depending on how and if you have developed any alternative scenarios. If this is the case then you may have a further section to compare the different results. Alternately, these may all become subsections within the model combination subsection.
You will likely also need a section for the site ranking process (solar radiation, views, comparison of model values and the like).
If ridges were used in the fire model then you need to give the source of the data or explain how they were derived.
Flow charts are very useful. You should improve on those given in the lab handouts.
Recommendations
This is analogous to the conclusions section from a normal scientific report. It does not need to be long. Fundamentally you need to say something along the lines of: “On the basis of our analyses, we recommend the university further investigate locations A, C and E (fig 10). Locations B and D are likely to be too costly to use, or may be overrun by cephalopods in wetter years.” Your conclusions will be somewhat longer than this, though.
References
See comments above.
Appendices
These are for large dumps of material that are not essential for the main text. Most of the figures should be in the main body of the report.