代做BEES1041 Exploring the Natural World代写Processing

2024.10.29 - 首页 >> Algorithm 算法

BEES1041 Exploring the Natural World

Communicating Results Report

Details

• The report is due by 17:00 on Friday 1 November (week 8).

• The report is worth 30% of the course assessment.

• The maximum length of the report is 2000 words. This includes all text presented in

titles, paragraphs, references, tables, and figures.

• Give your report clear sub-headings for each section.

Description

Figures

• A map showing the locations of the trees measured and the location of the Ausplot

Forests sites.

• Two scatterplots of tree DBH and height, with linear regression models showing how well DBH can be predicted from height. This is different to the models created in the computer exercise, which modelled height from DBH. One scatterplot and modelshould show the trees measured by the class, and the other should show the Ausplot Forests trees. Only the Eucalyptus grandis trees should be used, to make the comparison fair.

• A scatterplot of measured tree heights and lidar derived tree heights, with a root mean

square error (RMSE) value to indicate their similarity.

• Make sure you give each figure a clear caption, and that you refer to each of them in the text.

Table

• A table listing the number of samples, slope, intercept and R2 of the linear models

presented in the scatter plots.

• Make sure you give the table a clear caption, and that you refer to it in the text.

Text

• There needs to be short sections of text describing the results shown in the figures, and tables.

• The text should be organized under sub-headings that describe the different aspects of data analysis presented.

• Make sure that text addresses the two questions outlined in the introduction that I have written.

• References are not essential for a results section of a report, but if you do use some

make sure you are consistent and include a references list at the end.

Marking Rubric

Grade

Defining characteristics

(85-100%)

A report of a very high standard, well beyond that expected.

• Very high standard of written expression.

• Professional looking with a very clear structure.

• Excellent map, graphs and table.

• Demonstrated significant original insight into the topic.

high

(80-84)

Very little wrong with the report: not perfect due to one minor fault/issue.

• High standard of written expression.

• Good looking report with a clear structure.

• Good map, graphs and table.

• Demonstrated significant insight into the topic.

low

(75-79)

A very good report with demonstrated insights into the overall theme.

• Well written.

• A clear structure.

• Good map, graphs and table.

• Demonstrated insight into the topic.

high

(70-75)

A good report with occasional insights.

• Generally, well written.

• Clearly structured report.

• Clear presented map, graphs and table.

• Demonstrated occasional insight into the topic.

low

(65-69)

An adequate assignment but relatively simplistic.

• Mostly well written but with some issues.

• Simple report structure, with some minor problems.

• Map, graphs and table maybe missing some elements.

• Little creativity demonstrated in addressing the topic.

high

(58-64)

A poor report showing minimal effort and insight and some errors.

• Simplistically written with some issues.

• Report structure missing elements.

• Map, graphs and table maybe missing some elements and contain some errors.

• Limited depth of understanding of basic concepts.

low

(50-57)

A poor report with several errors showing minimal effort and insight.

• Simplistically written with several issues.

• Report structure missing several elements.

• Map, graphs and table maybe problematic or missing.

• Poor understanding of basic concepts.

(0-49)

The report did not fulfill the requirements in a meaningful way, or it was submitted late.

Frequently Asked Questions

What should you include in your results section?

You need to write the results section for a scientific report, presenting the tree measurement data made by the class. It should complement the introduction and methods sections that I have written, but you don’t need to repeat any of it, or include my figures. It needs to include your own figures, such as maps and graphs, a table and written text, and should be organised with some subheadings. Many people think that the figures and tables are the most important part of results. However, a scientific report or paper needs to describe the results in text as well. There doesn’t need to be lots of text, but there must be enough to give the writing some structure and put the figures into context. Some important ideas to think about:

• If you write too much text, a results section can be very repetitive, with the same data presented in the figures and tables and text. The text needs to describe what you think are the important results, but it should not try and describe all results.

• It is good to write your report without looking at the figures, to make sure the text

flows well, and there is a consistent message. Then refer to the figures by inserting

them into the relevant places within brackets. You don’t need to write, “Figure 4

shows that there is a close relationship between DBH and height.”. It is better to write “There is a close relationship between DBH and height (Figure 4)” .

• Use subheadings to organise the results. These might be different analyses, different datasets, different questions, or some other groups. If you like, you can describe how you have organised them in a short introduction. Some people like to include flow charts showing how separate data analyses are related.

• Make sure everything has a reason to be included. If you present some results in a

figure, then it must help answer the research questions and you should explain how it does this in your text.

Should it have some discussion as well?

Many people think it is a good idea to combine results and discussion together, but they are quite separate components. Results should only present what it is you have found, while a discussion section should compare these findings with what others have found, and sometimes integrate different results together. For this report, you don’t need to write a discussion, which requires reviewing literature, and is beyond the scope. But you should include some short statements about how the results help answer the research questions.

Should you write a conclusion?

A conclusion comes after the discussion and is used as a big picture summary of the significance of the work. You don’t need to write a conclusion section, but you can include a concluding statement at the end of the results, if you like.

What is the penalty for going over the word limit?

The penalty is proportional to how far you go over the word limit. Small amounts (<10%, 200 words) will lose 3 marks (10%), while large amounts (>50%, 1000 words) will lose 15 marks (50%). Make sure if you have a reference list it is included in the limit. It is very important to learn how to edit your work to make it concise. Make sure you read through your report several times and ask yourself if all the words and sentences are necessary.

Sometimes whole paragraphs can be deleted without losing any information.

Can you include extra figures?

Yes, you can include as many figures as you want. If you are thinking about using many more figures, consider organising them as sub-figures, which can make reading clearer.

Remember though, only include figures if they are relevant to the objectives and are referred to in the text, and make sure they all have clear captions.

Should you force regression models through zero?

Some people have asked me if it is better to force the regression models through zero, as we would expect trees to have zero height when they have zero DBH. This is an option when fitting a trend line to an Excel scatterplot, and can also be done in R. Although I understand the idea, and I may have used it in some of my work in the past, Ido not advise it in this case, for two reasons.

1. You are only looking at trees greater than 2 m inheight, so the model is never

predicting when DBH is zero.

2. Keep in mind that these small trees may grow in a different rate, with anon-linear

relationship between DBH and height.

The biomass model of Paul et al (2016) that I presented in the report introduction does not set the intercept to zero. If you google “when should you force regression through zero?” you will find many webpages where statisticians explain that in most cases, you shouldn’t.

What software can you use for the figures?

You can use any software to make the figures and conduct the analysis. I think it is important for you to learn some coding, but it is not always required. Make sure your figures are as clear as possible, so don’t use strange fonts and crazy colours.

Why do I want you to include analysis of the Ausplot trees?

You are able to investigate the questions asked in the introduction using the class tree data and the Sydney lidardata. So, you might be wondering why I have asked you to also include some analysis of the Ausplot Forests tree data? The main reason is so you can demonstrate that the model predicting DBH from height is only relevant to the types of trees measured, as the relationship between DBH and height is different for the Ausplot trees. Also, you might like to report whether the linear models are appropriate.

When is the report due?

The deadline is 5pm on Friday 1 November. Late work will be penalised by 5% of the value of the assignment per day. After 5 late days the work will be given a value of 0%. A 3-day extension is available without documentation athttps://specialconsideration.unsw.edu.au/

The following pages are the introduction and methods sections.

You need to write a results section to follow from these sections.

You should not include these sections in your report.

Comparing tree height to diameter at breast height and airborne lidar

Introduction

The biomass of trees are an important component of climate and carbon accounting models across the world (Paul et al. 2016). Measuring tree above ground biomass (AGB) requires destructive sampling (the removal of the tree), so it is common to model AGB from the stem diameter at breast height (DBH), which is 130 cm above the ground. These models can have very little uncertainty, as the relationships between AGB and DBH can be very strong. For example, Paul et al. (2016) developed a model (equation 1) of AGB from DBH for single stemmed Eucalypts (Eucalyptus, Corymbia and Angophora species), which achieved an R2 of 0.97 (Figure 1).

In(AGB) = 2.375 × ln(DBH) – 2.016 Equation 1

Figure 1. The relationship between diameter at breast height (DBH) and above ground biomass (AGB) of 6004 single stemmed eucalypt trees sampled from 225 sites across Australia, from Paul et al. (2016). The point symbols separate Eucalyptus species (FEuc_A) from Corymbia/Angophora species (FEuc_B).

The need to monitor tree biomass has led to the development of remote sensing methods to map different structural attributes of trees, such as cover and height. Airborne lidar surveys can create accurate canopy height models (CHM) as the laser pulses reflect off tree canopy elements, but also penetrate tree canopy gaps to reflect off the ground surface (Vauhkonen et al. 2014). Lidar cannot directly measureDBH, but models have been developed to predict DBH from lidar derived height in some forests (Bouvier et al. 2015), though there are few studies in Australia. The study presented here was designed to rectify this, through investigating the ability of airborne lidar surveys to model DBH for trees in Centennial Park, Sydney. This required answering two separate questions:

(1) How well can DBH be predicted from tree height?

(2) How accurate are lidar tree heights?

Methods

In September 2024 students in the UNSW course BEES1041 selected trees in Centennial

Park, recording their location, and measuring their DBH and height. Location (latitude and longitude) was determined using various mobile applications using the Global Navigation

Satellite System (GNSS). DBH was measured to the nearest centimetre using tape measures.

Height was measured using angles and distances to the treetop and tree base. All tree measurements were collated into a single dataset.

Airborne lidar data was downloaded from the Australian Government’s online portal, at

https://elevation.fsdf.org.au/. Lidar point clouds were acquired by New South Wales Spatial Services in May 2020 (Figure 2). The data were processed following the methods of Fisher et

al. (2020) to create a digital elevation model (DEM) and a maximum tree height model, or

canopy height model (CHM), at 1 m resolution. Lidar derived tree heights were determined

for each field measured tree by identifying the maximum height pixel value surrounding each location point, using a buffer distance of 5 m.

Figure 2. Lidar tree height over Centennial Park, Sydney, from May 2020. Note that some buildings are showing as false trees.

References

Bouvier, M. et al. (2015). Generalizing predictive models of forest inventory attributes using an area-based approach with airborne LiDAR data. Remote Sensing of Environment, 156, 322-334,http://dx.doi.org/10.1016/j.rse.2014.10.004

Fisher, A. et al. (2020). Modelling canopy gap probability, foliage projective cover and crown projective cover from airborne lidar metrics in Australian forests and

woodlands. Remote Sensing of Environment, 237,

http://dx.doi.org/10.1016/j.rse.2019.111520

Paul, K.I. et al. (2016). Testing the generality of above-ground biomass allometry across plant functional types at the continent scale. Glob Chang Biol, 22, 2106-2124,

http://dx.doi.org/10.1111/gcb.13201

Vauhkonen, J. et al. (2014). Introduction to Forestry Applications of Airborne Laser Scanning. In Maltamo et al. (Eds.), Forestry Applications of Airborne Laser Scanning: Concepts and Case Studies (pp. 1-16). Dordrecht: Springer.