I must admit that this part of the series is the hardest to write and the more I read on the topic, the more I understand the problem, but more importantly, the less I can think of what the solution is.
This part of the series concentrates on structural uncertainty and what this means for the model world and how we might deal with this. In the hydrological science there is currently an active debate on the topic, with different groups of people approaching the issue from different ends.
I will continue to use the linear regression example here, but I think the discussion actually might go beyond this model. The problem is that for the linear regression model we can actually mathematically prove that the line of best fit found by minimising the sum squared of errors is indeed the optimal and mean fit and explains the most of the variation in the points. With more complex models, this is not so simple.
However, I will start with writing what I originally designed and then I will write the disclaimers.

0 comments | Read more...

I finally managed to pick myself up to write another blog post. There are no apologies this time, just never got to it. So I want to pick up where I left off last time, discussing hydrological uncertainty.
The last thing I discussed was how the first uncertainty is in the actual fit of the model, or in other words, the model does not accurately predict all of the observed data, it predicts in fact the mean linear trend in the data.
Krzysztofowizc (1999) describes three main sources of uncertainty: uncertainty in the input data, uncertainty in the model structure and uncertainty in the output data. In this blogpost, I will concentrate just on the input and output uncertainty, my next blog post (and hopefully soon after this one) will deal with model uncertainty.

0 comments | Read more...

I want to write a series of blog post to try and explain uncertainty in hydrological models. Uncertainty is now a mainstream issue in the hydrological science and a growing number of publications are produced each year: between 2000 and 2013 there were 617 publications in water resources using the search term: Hydrology AND uncertainty AND model via Web of Science. The number of publications has increased each year from about 10 in 2000 to more than 90 in 2012. And it is not only in the academic world, the new groundwater modelling guidelines from the National Water Commission dedicates a whole chapter to uncertainty and CSIRO reports the uncertainty in the river modelling in detail.
But, if I read different reports, articles and discuss uncertainty, then it also a bit of a fashion, talking about uncertainty is similar to talking about climate change or the carbon tax. In other words there seems to be a lot of uncertainty about uncertainty. So I want to try to create some more certainty, and specifically about the sources of uncertainty in hydrological modelling.

0 comments | Read more...

Finally some rain fell this week. It has been very, very dry over August and the beginning of September, so I was very happy to get wet cycling to work. It was good rain too, not just a splatter, but a good bucket full. The interesting thing is that the Bureau of Meteorology in their seasonal outlook actually predicts a wet Spring for our part of the continent (look for the forecast periods August – October and September – November). So is the BOM all wrong?

0 comments | Read more...

When I was listening to Radio National on the way from Cootamundra to Goulburn, I was reminded that I wanted to write a blog post about the water accounting that the Bureau of Meteorology is currently developing. I might have written about this before, but I think it warrants highlighting. In particular, it is worth discussing what its current status seems to be, and more importantly what needs to be done and what it could mean for Australian water management.

0 comments | Read more...

Maybe it is pretentious to think that people are waiting for my blog posts, but I think I should just point this out. This blog is not as regular as it was in the past, and as it should be given the number of water issues that occur in Australia. This is not because I don’t want to, and so it really is a problem of affluence.

0 comments | Read more...

There are two random articles that I recently read that I found interesting but also disturbing. I think this blogpost is again about science and academic communication and how this can be interpreted, because both of the articles I read dealt with this topic. So it is more about the public perception about academic writing. Academics are expected to provide advice on issues that society faces, but it is interesting to note how this is done and how this is possibly perceived.

0 comments | Read more...

The Authors

  • Willem (Hydrology Research Laboratory)

About the Blog

Aimed at generating discussion on water research and water management in Australia
More

Other blogs

Water droplets Mike Young's comments on water (Droplets)

Pannell discussions David Pannell's discussions

Water Recycling in Australia Stuart Kahn's blog

Okham's Razor Clare Snow's blog