Creating disease risk map in QGIS? [closed]

Question

I would like to create a risk map using RFE Data and Tiffs but now I realised that I actually have no idea how.

I have dowloaded .Tiff(Region), .hdr and .bil(Rainfall/RFE) files from the USGS page, after that I georeferenced the .tiff and the .bil data to see in which part of the Region there is the most Rainfall( It's about malaria which has the highest risk in regions with the highest RFE.)

What I now want to do is to create a map that is clearly showing the regions where the risk is at its highest and a way to visualize it as good as possible.

1. 

2. 

Answer 1

The simplest way to run your analysis (without trying to define your actual inputs) is to do some basic raster algebra.

Have a go with the raster calculator: https://docs.qgis.org/2.8/en/docs/user_manual/working_with_raster/raster_calculator.html

You can build up some other vector layers too, such as 'distance to waterbody' as a distance raster.

To combine them, look into Multi-Criteria Decision Analysis (MCDA) https://en.wikipedia.org/wiki/Multiple-criteria_decision_analysis

The basic idea is that, say you have three rasters R1, R2 and R3 and for each cell (and the rasters should have the same extent and cell size) you're calculating an index by saying:

IND = R1 + R2 + R3

You can also weight this, by defining a weight for each raster, and combining that, so:

IND = (W1 * R1) + (W2 * R2) + (W3 * R3)

Additionally, it is normal to scale values for each raster so that they are between 1 and 0, you can do this by normalising each raster. If you're normalising then ensure your weights sum to 1, differences give you prioritisation of each factor.

Finally, you can have 'hard' factors, so, you might exclude the ocean from your study. Get a raster of your ocean, and classify that as 0 = ocean and 1 = not ocean. If you multiply your IND by that OCEAN layer, you'll only have values in the cells that aren't ocean.

Answer 2

When modelling the distributions of species, may they be animals or plants, fuzzy overlays are typically used. For an introduction into habitat suitability modelling, I suggest you take a look at GISLounge - Overview of Fuzzy Logic Site Selection in GIS. As Michael has already pointed out correctly, you will need more than one criterion (in your case rainfall) to accurately model habitats. Just from having a quick look at Mosqito World - Mosquito Habitats, I could already determine that land cover, water bodies as well as temperature should also go into your calculations.

Answer 3

I'm not specialist in diseases but have to deal sometimes species abundance distributions (SAD) with co-variables, which is methodically similar to the to disease mapping (see topic 5 at the end).

Risk mapping for diseases is a common task in science and statistics. The mapping context (in terms of making maps) is not so in the technical focus, because scientists, physician's and epidemiologist using terms like co-variable and location, which is more an address in the statistical/model feature space than things like x,y,z. I think for the malaria disease you can find a lot of statistical and mapping tools if you search for the buzzwords "risk mapping malaria co-variables". If you want a fast FOSS answers you can use additionally "r-package" for the statistical and the mapping context. Here some promising results from the WEB:

1. Topographic models for predicting malaria vector breeding habitats: potential tools for vector control managers Jephtha C Nmor, Toshihiko Sunahara, Kensuke Goto, Kyoko Futami, George Sonye, Peter Akweywa, Gabriel Dida and Noboru Minakawa http://www.parasitesandvectors.com/content/6/1/14
2. Mapping malaria risk using environmental and anthropic variables; Mauricio Edilberto Rincón-RomeroI; Julián Esteban Londoño II
3. Developing a spatial-statistical model and map of historical malaria prevalence in Botswana using a staged variable selection procedure; Marlies H Craig,corresponding author1,Brian L Sharp,Musawenkosi LH Mabaso and Immo Kleinschmidt
4. Using Structured Additive Regression Models to Estimate Risk Factors of Malaria: Analysis of 2010 Malawi Malaria Indicator Survey Data; James Chirombo, Rachel Lowe, Lawrence Kazembe
5. SDMTools: Species Distribution Modelling Tools: Tools for processing data associated with species distribution modelling exercises https://cran.r-project.org/web/packages/SDMTools/index.html]