1.2. Geo-Information Science and Sustainable Land Management (SLM)
Please read /docs/gentle_gis_introduction/introducing_gis. For an
introduction to GIS and what the field entails. Next, read the basics of
/docs/gentle_gis_introduction/coordinate_reference_systems. The
further you read, the more detailed information you get, and this is very good
for your understanding (
). We will also be using a UTM projection ourselves in the
practical. However, don’t get too lost on this!
1.2.1. Functionality of GIS
Note
This section was adapted from the manual.
GIS systems are equipped with a large set of standard functions for the analysis of spatial data (e.g. for combining maps or buffering). However, users often require very specific analysis. It is impossible for general-purpose GIS producers to provide all the functionalities for all users. Simple programming languages may solve part of the problem. In this practical we will work with different ways of adding functionality.
Although GIS packages may not be equipped with the standard functions required by users, in some cases we can combine several standard functions to create new functionality. This procedure works for example when implementing simple parametric models. In this course we will use the raster calculator in combination with some other GIS functions to implement the Morgan, Morgan & Finney (MMF) model.
Most GIS software is extensible by a programming language such as R or Python, in which the user can implement new functionality. The programming languages are typically for the advanced user and are also often used to develop user interfaces. The Graphical User interface in QGIS provides a /docs/user_manual/processing/modeler framework for designing and implementing geoprocessing models that can include tools, scripts, and data. Models are data flow diagrams that link together a series of tools and data to create advanced procedures and work flows. In this course, you will learn to use Graphical modeler to apply the MMF model.
1.2.2. Data quality and error propagation
Note
this section was adapted from the manual
Dealing with data in general or with spatial data specifically, we also have to deal with data quality and errors. Spatial data will always contain errors. The main question is whether the errors are acceptable. The latter depends on the use we will give our data. Dealing with complex forms of analysis we therefore have to study the effect of errors in our input data on the results of a particular analysis.
Although people tend to treat error as an embarrassing issue, it is very useful to make errors explicit. In ideal cases we do not represent absolute values as the outcome of an analysis, but we represent, for example, the risk of exceeding a certain value. Different techniques do exist to study error propagation. We can study the effect of changes on a particular attribute on our model outcome, but in reality, all the input parameters will include a certain degree of error. A more advanced method is called the Monte Carlo Simulation in which the distribution of all input parameters are determined. The simulation model will run many times and each time a new set of attributes will be drawn from their respective distribution in our model outcome.
During the practical you will be working on the creation of a Precipitation map based on a variable number of measurement points in which you will discuss and determine errors evident within the results.
New techniques have become available to avoid certain errors. Although many changes on the earth surface are continuous, we tend to classify characteristics and represent them as homogeneous objects on our maps. With fuzzy classifications, we are able to capture more of this variability. Instead of classifying a certain position to a certain class it is expressed as a continuous membership function having values between 0 and 1. A position can be a member of overlapping classes to different degrees. However, very few GIS packages already provide capabilities to deal with fuzzy classifications.
Finally, many characteristics of the earth surface are not constant as they vary temporally. We monitor these variables over time and introduce a new fourth dimension to our database: time. GIS that are capable to deal with this fourth dimension of time are called temporal GIS.