Lines in the Ice

Today I will talk about a fantastic exhibition I saw in British Library, called ‘Seeking the Northwest Passage’. It presented the history of the Arctic exploration from the Western perspective.

20150314_141545

The Europeans started looking for the fabled Northwest Passage after they became aware of the existence of America (I really don’t like the idea of ‘discovering’ America – it was there and it had thriving civilizations way before we put our hands on it). They suspected there was a connection between the Atlantic and the Pacific and so the expeditions to the North begun.

Obviously, I won’t summarize 500 years of Arctic exploration here – if you’re interested, the Internet is full of information and you can read about the exhibition here. But I’d like to share with you the main three points I took away with me from the day in British Library:

  • When you’re venturing into the unknown, the lack of correct visualization of the world is a big problem. All maps before mid-18th century were based on the assumption that sea can’t freeze. With this in mind, imagine the explorers’ surprise when they reached great masses of ice where they were expecting water. The realization that the scientists and geographers of the time didn’t get their facts quite right cost many, many lives (plus, plenty of time and money).
  • The change in the attitude towards the native people of the Arctic, the Inuit, was striking. For a long time the explorers went North to kind of ‘conquer the savages’. It took decades of unsuccessful endeavours to stop and think: “Wait a minute, if they live here it surely means they know how to be successful in the Arctic conditions – maybe we could learn from them?” Unfortunately, Europeans (in particular the Imperial British) those days were anything but humble. I really want to believe that now, in 21st century, the understanding of the Inuit culture and global effects of the actions we take (or not) in Europe is much better. However, while the Arctic has unceasingly been a subject of world’s politics – many countries tried to claim it, from the UK to the Soviet Union to Canada – the Inuit are still largely ignored.
  • The main motivation for finding the Northwest passage in 16th century was the idea of a quicker and presumably easier trading route to the Orient. Nowadays, in the age of climate change, the interest is still the same – people already calculated how much time, fuel and money melting Arctic would save the trade. One would have thought that as humanity we are advancing, but on the other hand – some things will never change.

The story of the Arctic exploration left we with many questions. Is the legacy of Europe’s colonial/imperial past bigger than we thought? How did ‘the conquering’ of unknown lands and depriving them of their wealth shape the world as we see it today? If the Northwest Passage is completely ice-free due to climate change, what will be more important to the global community – trade or the native people? How much has change in the past centuries, and how much has remained the same?

Advertisements

How to build an environmental model?

Today I am going to tell you the story of how I became an environmental modeller.

Once upon a time… which was four years ago, I did  a fantastic project in Mallorca, Spain. It was a part of my BSc Environmental Geography course at UCL, and it was my first adventure with modelling of any kind. Little did I know that that adventure would inspire me to choose modelling as a career.

The aim of the project was to model the hydrodynamics of Alcudia Lagoon in Mallorca, and its relationship with the Alcudia Bay to which it was connected through a channel. You might ask yourself, why on earth would anyone want to do that? Well, to understand the lagoon’s behaviour (if only we could do that with people…) and to know how sensitive the whole system is to any changes or disturbances.

Below you can see satellite images of the lagoon.

mallorca

So, how do we build an environmental model? First of all, we had to gather the field data which would drive it (water level and bathymetry at the lagoon and the bay), and the field data we would use to test our model’s accuracy (flow speed at the lagoon’s channel). That was the fun part – we got to pontoon around the bay and sit on the beach, as you can see in the pictures below.

298652_10150365545019952_1002113810_n

Measuring the bathymetry of the lagoon and recording the GPS coordinates of the measurements.

303979_2485004638772_1513195548_n

The bathymetric measurements were tricky sometimes!

339134_2485001598696_358113985_o

Setting off at 6am to locate a water level meter in the Alcudia Bay.

390663_10150365543794952_269124292_n

Measuring the water flow speed with a propeller.

After three days of collecting data, the girls in my group and I were ready to start using the Surface Water Modelling System (SMS), a programme allowing to create hydrodynamic models. The specific model we used was RMA2 model – a computer code for two-dimensional simulations. But first things first – every modeller needs to start with a conceptual model, to get their head around the system that is being modelled. Conceptualization could be seen as the foundations, or the base, or the frame of a model. In this project, we needed to delineate the geometry of the lagoon based on a satellite image – that was the spatial component of our model. Then, the SMS created a computational mesh based on out delineation, and it looked like this:

mesh

Once the mesh was ready, we used the water level and bathymetry to fill it in with information. It’s important to note that we only used the data collected on Day 1 – we needed the data from the other days for later.

Most of models are developed to predict a system’s behaviour, so it is absolutely essential to be able to verify their predictive power. How do we do it? It is actually really simple – we let the model simulate something we’re interested in, and then compare the results with the actual field data. In the case of our lagoon model, we asked the SMS programme to simulate water flow speed at the outlet channel, in the same spot where we collected the data. Through  calibration, which is a process of playing around with some of the model parameters (obviously, within reason), we found the best fit between modelled and observed data. In this project we fiddled with eddy viscosity and friction at the bottom of the channel. Both parameters are really hard to measure in real life, so varying them to find the best fit is a perfectly legitimate thing to do!

Our best model looked like this, and believe it or not – it was a pretty good result:

Calibration Graph

But then, if the data fits your model once, it doesn’t necessarily mean it will always be so. That was why we need to validate the model against another set of data, and that was where the data from the other days came in. We simply run the model again, but using the data from Day 3, and it looked like this (a little bit worse…):

Validation Graph

Overall, our model proved to be quite efficient! We explored the hydrodynamics of the lagoon and we even simulated what would happen if the sea level rose by 30cm, as suggested in one of the IPCC reports. Here’s what we found:

-the lagoon is mainly affected by the seiches and tides occurring in the bay;

– the water level oscillations in the lagoon decrease with the distance from the sea, which means the channel might be a ‘filter’ for the variations in sea level;

-water flow velocity depends on the sea level and it is lower when the tides are peaking;

-sea level rise could lead to decreased flow velocities, and that could lead to sedimentation of the channel and potentially cutting the lagoon off.

So, that was my first environmental model ever and I absolutely loved that project – from collecting the data to actually making sense of it. For those of you who haven’t had much idea of how to build an environmental model, I hope this post explains it all step by step! And I hope I helped you realise how fascinating environmental models are!

London looks pretty in data

A new book, ‘The Information Capital’, is coming out soon. It gathers lots of information about London – all compressed into telling images. More information here: http://www.londonist.com/2014/10/the-information-capital-london-looks-pretty-in-data.php

Generation-Rent-p64-65

Median flat prices across London. White (mostly central) areas show no properties under £250,000. Then, the lighter the colour the lower the median flat price. Yellow bands show properties selling for £85,000.

Visualising the present, visualising the future: a brief introduction to environmental models

I am an environmental modeller and to me environmental modelling is, unsurprisingly, one of the most important, useful and efficient ways to visualise the world. Throughout the blog I will be presenting many of the modelling projects I have done in the past, and this is why I think that a brief introduction to environmental models is essential. In this short note I will try to explain what modelling is, why we need it and what challenges it involves.

We, humans, are an incredibly curious species, and that innate curiosity is reflected in our history. The world as we know it today is a result of thousands of years of exploration and innovation. The development of humanity has been driven by our tendency to question the reality and never cease to look for answers. The two most commonly posed questions that drive scientific advancement are: “why?” and “what if…?”.

Finding answers to these questions used to be rather costly and time-consuming; however, people learned to use their observation or knowledge to replace direct experimentation by simply simulating reality. Nowadays models are present in virtually every aspect of our lives – they are used by, amongst others, scientists, policy makers, economists, engineers, teachers or architects.

From now on I will focus on environmental modelling, as that is my main area of interest. Environmental models have two main purposes: helping understand the behaviour of the environment (answering the question “why?”) or predicting it (answering the question “what if…?”). In other words, they give us the power to simulate environmental processes and predict their possible outcomes. They play a crucial role in overcoming the constraints of time and space: they provide a means of quickly extrapolating from existing measurements to remote, inaccessible areas where it’s difficult to collect data; and they allow us to foresee the future (or at least attempt to). They are also invaluable tools that can compensate for the shortcomings of current measurement techniques and, by quantitative extrapolation, provide the data for the variables we are not yet able to physically measure. This is the reason why, over the years, environmental modelling has become an indispensable part of policy-making, monitoring and experimental research.

Unfortunately an intrinsic part of any kind of modelling is uncertainty. It cannot be forgotten that models are just simplified representations of reality and are often based on incomplete data or approximate estimates of parameter values. One of the biggest and most important challenges in the future of environmental modelling is reducing the uncertainty. This will involve improving data acquisition techniques – no model can be more accurate than the data that it is based on. Another way of improving a model’s performance would be the combination of modelling approaches (which I will describe soon) – merging different modelling strategies can be very successful. Finally, in order to be able to tackle complicated, large-scale environmental problems, such as climate change, a universal modelling interface is necessary. The ability of putting the outputs of different models together would be a great step forward. The importance of interconnections between the components of the global system is undeniable, and should be included in environmental modelling to the greatest extent possible.

Visualising the world

A few years of experience with environmental models taught me one thing: the world is way too complex for us to entirely comprehend it. However, I have to give us, humans, the credit for tirelessly trying to do so. As an incredibly curious species, we have been gathering enormous amounts of data and we have been looking for patterns within them that would help us understand what is going on around us. This strive for better understanding does not solely aim to satisfy the curiosity – it is also necessary for decision making in almost all areas of our lives. It is very simple – the more information we have, the better decisions we can make. Information allows us to assess the situation, weigh the pros and cons, and predict the consequences of an (in)action.

We would think then that the more data we have, the better; and that is where we come across problems. In this place we have to acknowledge that the data we gather on the environment is often of quantitative nature – at first, we describe the observed processes and relationships conceptually, but then we try to depict them in the form of numbers and equations. And although our brains are very sophisticated machines, they might not be able to quickly process countless rows and arrays of numbers. In this case more information (data) means more confusion, and way more time required to make any sense out of it.

So, what can be done about this? Well, an important thing to remember is that humans are visual animals. Sight is our major sensory means of gathering information about the world, so for easier and quicker understanding of complicated data we need to make it graphical. And, ideally, aesthetically pleasant. This allows swift picking up of patterns and lets us focus only on relevant information. In other words, we can easily get an impression of what a load of data means – no detailed, time-consuming and effort-inducing analysis required.

Undoubtedly, visualisation of data is crucial for efficient communication of information; however, due to that incredible efficiency, it is very powerful. And we all know that with great power comes great responsibility. Presenting the data in such a manageable form requires compromise and simplification. This consequently gives lots of room for selectivity and misleading the recipients of the presented information.  That is why while visualising our environmental data we have to be very careful – we need to be aware that by presenting information in a visual form, we create a lens through which others see the world.

ilustracja 1