Brontekst bekijken van Modelling Land Use Changes

{{Infobox|data2=Van Rompaey Anton|data3=Lectures and practicals|data4=Written exam and assignments|data6=6|header1=Courses and exams|header5=Background|headerstyle=background:lightgrey|label2=Prof|label3=Courses|label4=Examination|label6=Credits|label7=When?|data7=2nd term|label8=ECTS|data8=[https://onderwijsaanbod.kuleuven.be/syllabi/e/G0I83AE.htm Link]|title=Course Information}}
The exam (10/20) evaluates the insight and knowledge of the themes that were discussed in the course and the practicals. Students should write a report of their home assignments (10/20). Not taking part in the practicals results in a failure on the exam.

The exam used to be oral and worth 16/20 points (before 2021).

== 2021 ==
''(written)''

=== June ===

# Define a set of push and pull  functions for urban (pixel X). Explain the method. (9x9 grid was given with Urban, Arable and Industry pixels)
# How can you prove that a change in NDVI is the result of human influence (land use changes). A table was given with NDVI and rainfall values for the period 2005-2020.
# Assign the pixels to forest or arable using MOLA and IMLA, Discuss potential differences. (2x2) grid was given with probabilities for forest and arable land for each pixel.
# Explain briefly:
## Maximum Likelihood calibration
## Stochastic pertubation
## Error due to location

== 2020 ==

=== June ===

# Calculate the ROC value. A 3x3 grid of probability values is given, and an observed pattern of forest is also given. Explain the procedure.
# Develop utility functions for agents that lead to the formation of this system. Figure 6 from the paper from Parker and Meretsky was given.
# How are the parameters of a logistic regression calibrated?
# Explain briefly:
## Elasticity
## MOLA
## Null resolution
# Explain the graph (see figure below UP-SH)
# Assign land use category in 2*2 grids using IMLA and MOLA. Probability of each land use is given

== 2019 ==

=== June ===

# Calculate the ROC value. A 3x3 grid of probability values is given, and an observed pattern of forest is also given.
# Calculate the probability of urban for location x in a simple 4*4 LUmap. This was a grid with urban, forest, arable and grassland. Ruleset for CA was given (Fig. 4 in Poelmand et al., 2010).
# Assign 1 forest and 3 arable pixels in the following 2x2 grid, using IMLA and MOLA. A grid with probabilities for forest and arable was given. Explain and give possible differences.
# What is the null resolution? Give an example.
# What do we learn from these graphs?

# Negative externalities cause a fragmented land use pattern, explain with examples
# How can error due to location be assessed?
# How are the parameters of a logistic regression calibrated? Give an example.

No questions about the assignments

== 2018 ==
[[Bestand:Graphs from exam modelling land use changes (2020,2019,2018).png|miniatuur|Graphs from exam modelling land use changes (2018-2020)]]

=== June ===

# What do we learn from these graphs?
# What does this relationship mean? How can it be used in a GIS?
# Explain with an example:
## Land Use Elasticity
## Stochastic Perturbation
## Negative feedback of land use
# Question about the assignments. (He just wanted to know if you were a "freerider" or did really cooperate in the task).
## What neighbourhood was used?
## How was the model validated?

== 2016 ==

=== June ===
The exam was closed book:

# Calculate the ROC value. A 3x3 grid of probability values is given, and an observed pattern of forest is also given.
# Calculate the probability of urban for location x in a simple 4*4 LUmap. This was a grid with urban, forest, arable and grassland. Ruleset for CA was given (Fig. 4 in Poelmand et al., 2010).
# Assign 1 forest and 3 arable pixels in the following 2x2 grid, using IMLA and MOLA. A grid with probabilities for forest and arable was given. Explain and give possible differences.
# What is the null resolution? Give an example.
# Assignments. (The question was on the exampaper but he did not ask anything)

# Derive push pull function for urban at the location of pixel x, figure was given. So you must be able to define the weights for different distances.
# What is the error of quantity and location on this image. What is the null resolution?
# How can you address human impact on a vegetation, in other words how can you distinguis these factors from rainfall? NDVI values given and precipitation values, you must be able to explain method from the paper.
# Explain: 
## What is land use elasticity?
## What is stochastic pertubation?

== 2015 ==

=== June ===

# Given: NIMBY figure of agent-based modelling. Explain the pattern. How are the profits calculated?
# Given: 9x9 grid with urban, agricultural and industrial land use. Calculate sets of weights for location x in that grid.
# How are the parameters of a logistic regression calibrated? Give an example.
# What is the null resolution? Give an example.
# Assignments (easy questions like: How does MOLA work? Did you calculate the weights for the CA method or were they given? Did you used logistic regression or CA for this assignment? etc)

# Explain figure 3 from paper Clarke&Gaydos about CA
# How to calibrate parameters in logistic regression?
# Explain 'null resolution' and 'land elasticity'
# Negative externalities cause a fragmented land use pattern, explain with examples
# How can error due to location be assessed?
# Assignment question (how did you perform sampling? How did you determine which parameters were significant in logreg?)

== 2014 ==

=== June ===

# Give utility functions for the following 'problem':
## Rich people like living on high elevation and they have a car, so they want to be close to roads, but they don't need to live close to the city centre.
## Middle class likes being close to the city centre and close to roads to take public transport or car.
## Poor people can't afford a car, so they just want to live as close as possible to the city centre.
## People from a higher class hate it to live in the neighbourhood of people of a higher class. Use a map to explain how this utility function can help allocating land use
# Figure from the paper of agent based models (urban sprawl: negative effect of urban on urban and from urban to agriculture and vice versa). Explain.
# IMLA and MOLA: allocate land use in a 2*2 grid on both ways. Are there differences?
# Question assignment (what software did you use for that?)

# Calculate the ROC value. A 3x3 grid of probability values is given, and an observed pattern of forest is also given.
# Calculate the probability of urban for location x from the following grid. This was a grid with urban, forest, arable and grassland. Location x was marked. A graph with the weight's was given. This was the same graph as in the last slide of Cellular Automata slideshow.
# Assign one forest and three arable in the following 2x2 grid, using IMLA and MOLA. A grid with probabilities for forest and arable was given. Explain and give possible differences. (Using IMLA, you end up with an oscillation and the necessary amounts are never reached. Question on oral: how would you solve that problem?)
# Assignments. He randomly picks one of your assignments and scrolls through.

# Explain and apply the principle of push and pull factors on pixel x of a 9x9 grid.
# Given this data set of NDVI values and precipitation values. Is there human interference in this system? If so when and how can this be seen?
# Explain: Land use elasticity, stochastic perturbation and negative feedback in land use systems.
# Short discussion of random assignment. (He doesn't really know what they are about though)