Questions and Answers

0 Dislike

Alec Wild

Node7-8 what does the probability field mean?



I am trying to obtain the conditional probability of ash dispersal? I have set it to conditional and using the "average" selection for statistics. But I am unsure what the probability is related to? I am trying to retrieve the hazard map for the eruption pathway up to node7-8.



Report abuse

2 Responses

  1. 0 Like 0 Dislike

    Roberto Tonini

    Dear Alec,
    first of all, thank you for using our tool.

    In your case, at node 7-8, you are calculating the conditional (i.e., given a specific "eruptive setting") probability, in a given time frame DT, to overcome a set of discrete values of tephra load or thickness (I don't know which intensity measure you have selected for your study) for the target area that you have defined (a grid of spatial points). "Eruptive setting" means that you have selected location and size/style at previous nodes: so, given an eruption from a specific vent location and of a specific size/style, you are calculating the impact of tephra on your target area by calculating the probability that tephra could overcome a set of thresholds in each target point.

    This is usually represented by means of hazard curves, having on y-axis the average exceedance probability and on the x-axis the values of intensity. These hazard curves are calculated for each point of your target domain. When you select to plot the hazard map, you are cutting all these curves by fixing a value of exceeding probability (let's call it PTH). As consequence, on your hazard map, you will have the value of intensity having exceedance probability PTH in the time frame DT. In PyBetVH tool you can choose interactively the value of PTH (between 0 and 1).

    The BET models do something more concerning hazard curves. Through Bayesian inference they account for the uncertainty associated to the exceedance probability itself. This means that each value of probability in the hazard curve has an uncertainty (defined by a distribution of values). So, for each grid point, PyBetVH calculates a set of curves showing the degree of uncertainty of your estimation. When you chose the "average" statistic, you are choosing to plot the average curve of such set of curves. For practical usage, very often, people look at the average only, but it is important to have an idea of how large the error in the calculation can be.

    Looking at an example could result clearer: in Figure 1 of the PyBetVH paper (you can find it here you can see that there are 4 hazard curves plotted: the average and then the 10th, 50th and 90th percentiles, which give you an idea of the uncertainty associated to your probability estimation. In the same figure it is also explained how hazard and probability maps are obtained.

    I hope to have properly addressed your question. Please let me/us know if you need further help.
    Best wishes

    Reply Report abuse

    Please login to answer the question.

  2. 0 Like 0 Dislike

    Fernand Pol

    The doubts related to conditional probability of ash dispersal will be always make people confuse. There are lots of probabilities it will come with result. So we should be aware of all the possibilities and it will make our mind feeling stressed. As part of my job with dry cleaning service in Noida i am consulting so many people. They have also some doubts about this. 

    Reply Report abuse

    Please login to answer the question.