Waching daily May 29 2017

The project is financed by the European Research Council se llama 'Statistical Inference

for Remote Sensing Earth Observation Data Analysis'.

This project was born in 2015, rather recently

The field of research that includes this project but also other tasques inside

the group is the advanced prcessing of remote sensing images,

observation images of the Earth.

But not only observation of the Earth is formed by images that acquire

the sensors inside the satellites but aslo by measures that we take in the field,

in the oceans or in the atmosphere.

We try to develop efficient algorthms so that these models can provide

better predictions of temperature, of humidity or of the chlorophyll contents of the plants.

The project is based in using statistical techniques for the data treatment

of the observation of the Earth in general, so, let's say that we live in this intermediate world

between statistics or maths and the Earth observation or something more biological.

Oficially, been hired by the project right now there are three post docs

and two students, a part of Gustau who is PI; and we have to contract another student

but extraoficially the are far more people working behind because we collaborate

with a lot of people.

I'm a researcher of the project and a big part of my day I'm in front of

the board with my colleagues and then we sit and we make simulations to know

if our patterns are better to explain the information of the satellites.

That is, if they are better in predicting interesting variables. In other words: how it is

the Earth in places where the satellites are looking.

The group of research is an intermingling of many researchers who belong

to different departments of the University, from electronics, to optics, maths...

At this moment we can say that the project is divided in two big strands.

One is the one that improves algorithms automatically of prediction of climate variables, for

example the temperature, the humidity, the concentration of ozone in the atmosphere...

And also, on the other hand I think that the most significant part of the project is the fact of jumping

from some patterns of prediction to patters that try to explain things.

This project is important because not only we mix these two worlds, of the biology

and of statistics, but also it is important right now because in the observation of the Earth

there are a huge amount of information, a thousands of milions.

There's a lot of satellites above us taking pictures and all that information is very

hard to process and even harder if it is one person who has to look at them one by one.

Let's say that it is impossible, so these statistic methods extract that information in

automatic way and it help us to propose physical patterns but let's say that the

most innovative part of the project is the causality part.

The essential singularity is that we are trying to develop new mathematical

techniques that do not yet exist in literature

that try to deal with both main problems.

An essential problem right now is to trust the predictions that we are doing

for example for the temperature.

On the other harnd, many patterns are simply predictors, they estimate the climatic viariable,

the interest, but they don't explain why this variable exists.

Which are the tolerable levels or not.

Like this, what we are trying is to advance in more complicated patterns,

more complex that we call 'patterns of causal inference' which try to find links

between the variables and the mesures that we are seeing.

We are involved right now in much more risking initiatives, one of those is looking

back, not forward like the predictive factor of the patterns, but making anticausal

factors: try to predict the past on the basis of the present so that we could know

how we've arrived at this point.

Furthermore, there's another issue that we are interested in which is knowkn in the literature as

'planet bounderies' that are the limits of the Earth.

To know which are the multidimensional limits of the Earth.

To give a very clear example: maybe we could live in 2050 with 5 more

degrees in temperature, but could we live with 5 more degrees

and 5000 more millions of people living on the planet?