The scope of Environmental Science and Scientific Thought. From Thought-driven to Data-driven, from Critical Thinking to Data Management. (by Diego Fdez-Sevilla)
Pdf available at Researchgate: DOI: 10.13140/RG.2.1.2007.0161
Environmental science use to be a Thought driven, Knowledge based and Data handling dynamic scenario focused on finding creative and systematic ways of understanding and analysing our surroundings.
In project management, the term scope has two distinct uses- Project Scope and Product Scope. Scope involve getting information required to start a project, and the features the product would have that would meet its stakeholders requirements.
Nowadays, it seems that the scope in Environmental Science and the role of a scientist has moved from Thought-driven to Data-driven, from Creative and Critical Thinking to Data Management.
From Thought-driven to Data-driven.
At some point in time an Institution gathering environmental data faces a breakdown in communication with a weather station located in the slope of a mountain.
The call is made for two members to go by car to the location and evaluate the situation. Two colleagues from different departments agree to go together. Considering previous data from the location they expect to face snow and cold conditions at this part of the year. Both take winter cloths to be prepared and hire a 4×4 vehicle.
They travel to the location. At one point, they leave the main road, going over dirt mountain roads. The terrain becomes muddy on the way up to the location.
Once they arrive at location, they leave the car. There is not snow but it is cold. They walk to the parcel surrounded by a metallic fence where there should be a pole with different instruments attached.
When they arrive they see that the fence has been broken and the pole has been bent causing damage to some of the instruments. They look around for clues to understand what has happen and they see fur in the fence and some footprints. One of them points out that it looks like a bear has been the cause for the “trespassing”.
They take some pictures, pick up the instruments into the car for later repairing and leave the place.
When they arrive at their departments each one shares the experience with their office mates.
One of them describes the situation as a total crisis, showing the pictures pointing out the damage suffered by the instruments. All leading them to lose unrecoverable amount of data having to rely on averages from the nearest stations. Furthermore, ongoing research will have to be reassess.
The other one describes the situation as a good opportunity to study the evolution of the microclimate in the region where the location is. He starts to points out that, at some point moving uphill, he/she noticed that the snow line expected for this part of the year had retreated and that the road should be frozen instead of muddy at the altitude for the location of the station, and the time of the year. He/she also noticed that could recognize some species of plants around the area of the station which normally are associated with habitats that develop at lower levels in altitude. Furthermore, bears at this time of the year should have reduced their activity to a minimum and undergoing their “winter lethargy”. Enough data to start looking into some previous publications for further research. He/she even had the pictures to add into the study.
Science is undergoing a transformation which follows a pathway led by the term “data”. Scientists are no longer thinkers growing on knowledge looking for connections. They start to become data managers, data scientists relying on algorithms to understand the mechanisms driving the dynamics which have generated the genetic drift, black holes, gravity, energy and intelligence.
Sometimes, “data” is what it is being generated by your capacity for understanding. Other times, “data” is what allows you to understand. It might be a matter of discussion if “data” can teach how or what to understand.
Blinded by technology. From Critical Thinking to Data Management.
If a solar flare hits the Earth dropping down the functionality of our hi-tech based infrastructure, what would we be able to see, without hi-tech instruments? Well, that is a worry part of a sci-fi film and other implications might be more important for our daily routines. However, satellite data is becoming the primary source of information for most of the research projects aiming to explain the behaviour of our environment. So, what are the implications of relying the capacity of generating assumptions on satellite data?
The Satellite data is obtained and applied on environmental studies rely on two basic principles: the capacity of the material under study to emit or reflect any form of energy that our sensors can measure, and the “transparency” or “conductivity” of the medium through which such information travels from source to receiver.
Something as simple as changes in air quality due to air planes traffic, dense concentration of pollution, or the eruption of a volcano will interfere with the reception of any data from areas covered by the cloud formed by the aerosols and gasses released. And that would spread in the atmosphere affecting the accuracy of measurements in more locations.
In microscopy, the refraction index of the embedding medium used to contain the sample under study interferes with its visual quality, similarly, the composition of the atmosphere (the embedding medium of our environment) is related with the coefficient of transmission for the spectrum of electromagnetic wave lengths being studied. Similarly, also with the alteration of the composition of the atmosphere, in altitude and in longitude, there are other factors affecting the accuracy of the data received by satellites such as the atmospheric thickness and the angle of incidence between the source of energy in relation with the receiver (either passive or active sensors).
One example about the limitations of relying on satellite data to make assumptions comes from the SeaWiFS instrument aboard the SeaStar satellite which has been collecting ocean data since 1997. In theory, by monitoring the color of reflected light via satellite, scientists can determine how successfully plant life is photosynthesizing. What it is not so simple is to relate measurement of photosynthetic activity with measurement of successful growth. You can have trees with photosynthetic activity but low density tissue due to wood decay triggered by fungi (e.g. white or black rot). But also, growth not always means successful use of ambient carbon as we can see when we face the problem of eutrophication. Algae, like land plants, capture the sun’s energy and support the food web that leads to fish and shellfish. They occur in a size range from tiny microscopic cells floating in the water column (phytoplankton) to large mats of visible “macroalgae” that grow on bottom sediments. Algae may become harmful if they occur in an unnaturally high abundance or if they produce a toxin. A high abundance of algae can block sunlight to underwater bay grasses, consume oxygen in the water leading to fish kills, produce surface scum and odors, and interfere with the feeding of shellfish and other organisms that filter water to obtain their food. Some algal species can also produce chemicals that are toxic to humans and aquatic life.
Even considering all the potential behind the use of remote sensing thanks to satellite data, this methodology has to be approached cautiously understanding its limitations due to the restrictions of the technology in itself.
(update Oct 2015) A recent publication (October 30, 2015) points out that “degrading satellite sensors, not soot or dust, are responsible for the apparent decline in reflectivity of inland ice across northern Greenland.” Geophysical Research Letters, 2015; DOI: 10.1002/2015GL065912
Other example comes in the following article describing the limitations of the data and the processes associated to correct the bias incorporated by the data. Satellite data regarding the eutrophication response to human activities in the plateau lake Dianchi in China from 1974 to 2009)
Another misconception has been adopted from observed increasing measurements of atmospheric CO2 and its potential boost on photosynthetic activity. Ecosystem effects of increasing levels of atmospheric CO2 will depend on the nutrient status of specific forests. Increased forest production will occur where soils contain adequate nitrogen. In areas where nitrogen is limiting, elevated CO2 levels will not increase the growth of trees — even though photosynthesis may increase. Without sufficient nitrogen, the trees cannot use the additional CO2 for growth. The additional carbon is used by soil organisms and respired to the atmosphere. In addition to contributing to CO2 buildup in the atmosphere such changes in the soil foodweb, which controls nutrient availability for plants, could have long-term effects on ecosystem functioning.
Understanding how much it is being affected the capacity of natural systems to not only stabilize Carbon in structures, but also, to keep them inactive, changes completely the assumption of what we assume to behave as carbon sinks.
Technology is overtaking, and somehow overriding, human patterns of thought and capabilities for analysis and discovery.
Algorithms and digitalization are encapsulating the power of observation and “random” thinking. Random thinking is a key characteristic behind any discovery since involves the integration of all types of information (multi-sensorial, mathematical, emotional and multifactorial in time and space) into a multidimensional and multivariable state of mind and processing. Digitalization of information, as it happens with music (from analogue to digital), reduces the amount of information restricting the meaningfulness of the data to the limited rules which define and sustain its own singularity as a digital language. This situation makes the interpretation of data subject to restrictions, narrowing “perception”, data integration and mind and artificial data processing, into “only” that that this language is capable of transmitting.
As much as from planting apple trees you can not expect to obtain oranges applying digitalization and algorithms to describe the environment implies to restrict the outcome from a digital platform into a digital recreation of digital interactions. One of the potential limitations that I suspect comes from relying on algorithms is the lack of creativity by those to identify new variables playing a role, and their limited capacity to assess the implication of integrating new variables in the equation. As an example, the behaviour of our environment, and all the variables interacting and playing different roles, follow some patterns of inertial momentum which can be identified and recreated at some extent. However, there is a big lack of information about thresholds. Something like considering how far can you recreate with algorithms the changes coming into your life when you know that you are going to have a baby.
I believe that science in general, and environmental science in particular, base their functional power on perception. From there, interpretation and then assimilation. All those steps lead to understanding and potential discovery. Only by identifying the limitations of our perception of things we will understand the problems behind assimilating new concepts and our lack of understanding.
Environmental science is becoming a data driven organization. And therefore it carries the risk of becoming blind by technology. It is moving towards seeing things only through technological interpretation. And yet, there is a huge gap of knowledge addressing its limitations. Scientists are becoming data managers subjecting their creativity and capacity of perception to data processing and algorithms. Consequently, if the data does not show it or the algorithms do not replicate it, it does not exist.
There is a transformation in the scope of Environmental Science. Ranging from Micro to Macro, the focus of attention is being displaced from Thought-driven to Data-driven, from Critical Thinking to Data Management. And we might not realise what we are losing in the process.
It would be scary to find out one day that somebody has realised that our models and/or algorithms applied to filter the data released from satellite sensors, have carried a miscalculation and any previous assumptions based on them are wrong. When at the same time, all parts of our ecosystems have been sending signals strait to our senses which nobody knew how to interpret properly.
If in today’s time, with all the revolution in technology developments, we still are facing different description for the same reality from different scientific approaches about our environment (similarly as the tale of “the blind men and the elephant”) the conflict might come either from:
- restrictions in the methodology applied to retrieve data and processing, or/and,
- restrictions in the capacity of perception and interpretation applied in identifying the meaning of the data obtained.
Contact with Envisat was suddenly lost on 8 April 2012. Following rigorous attempts to re-establish contact and the investigation of failure scenarios, the end of the mission was declared on 9 May 2012.
We observe by-products and activity originated from what is happening in our environment; particles, gasses, land or water movement, species settlements and displacements, temperature, and wave radiation of different kinds. But, after the perception of their existence, understanding the mechanisms behind the generation of these “signals” is the real question. And at this point, satellites share some limitations with those from microscopes. Just because we can see things through their instruments, it does not mean that we fully understand what it is that we are looking at and what it is going on in front of our eyes. Which part of the reality are we observing? How much of the reality that we are observing comes from how the instruments are affected by side effects?, how much of the reality that we are observing comes from the limitations in our understanding? and, how much of the reality that we are observing comes from the limitations carried within the narrow selection for our focus point of attention?
Here I want to share my experience from applying optical and digital imaging to perform automated and none automated analyses identifying features in order to characterise samples of particles.
The subject of discussion was originated from the following question: Can Advance Microscopy Particle Analysis replace other techniques as a primary technique of Particle Distribution and analysis?
We can compare “Microscopy” with “Satellite Observation” keeping in mind the similarities and differences since both methods apply tools to observe and create images, and interpret their content and significance. Through microscopy I have studied patterns of particle deposition over different surfaces (tapes and filters of different material composition) as well as morphological and optical features applied to identify biological and none biological particles (transparency, refraction index, superficial roughness as well as the existence and morphology of pores and furrows).
There are several techniques which can be applied. Biological particles like Pollen grains, can be analysed by microscopy with white light using the setup of mechanical elements such as a condenser, a diaphragm, etc and variations in your point of focus to obtain optical optimization and look through different depths obtaining a 3D composition. Also we can polarize the bean of light or use dark-fields enhancing contrast with the background regulating the angle of incidence from the light bean. Furthermore, biological tissues have the unique particularity of fluorescence under certain wave lengths that makes them stand up out from the rest of the material in the sample.
Satellite observation, as well as microscopy, applies techniques which can be based on optical imagery or wavelength spectrum analyses. And as well as with microscopy, the measurements are affected by the conditions of the sample and the characteristics of the material under observation: water, land, ice and snow, gasses, plants and aerosols. And also, by the medium through which our information travels; the atmosphere: its thickness, its composition (gasses and aerosols) and thermodynamic conditions (e.g. thermal conductivity).
Filtering the relevant part of information from the interference of other sources which interact with the depth of field and the accuracy of the measurements is a challenge. And it can not be emmascarated by focusing our attention in prioritising making assumptions based on big quantities of data.
In today’s time, similarly as the tale of “the blind men and the elephant”, after all the evolution in technology developments, we still are facing different descriptions from different scientific approaches for the same realities about our environment.
There have been two conclusions which have become more frequent in published research:
- the amount of data is not enough to draw conclusions, and,
- there is a need for new theories to understand what to do with the increasing amounts of data.
A thought-driven mind is creative. Proposes a theory based on knowledge and then search for data to validate it. A data-driven mind is reactive. It looks at data and reacts to what the data tells.
Science and scientific thought is becoming “reactive” instead of “creative”.
A creative mind has restrictions defined by knowledge. Meanwhile a reactive mind has restrictions either from:
- the methodology applied to retrieve data and processing, or/and,
- restrictions in the capacity of perception and interpretation applied in identifying the meaning of the data obtained.
In our global society, the understanding of people demands, needs and personal likes and dislikes, have been started to be replace by numbers. Empathy, the use of our senses to gather information which would help us understand the behaviour of the people around us, is been replaced by statistics. So companies have began to treat people as numbers part of a market.
In science, similar situation is happening with our environment. Each part of the ecosystem is being categorised numerically and its behaviour relativised to “known” variables. Consequently, if the data does not show it or the algorithms do not replicate it, it does not exist.
We are at risk of reaching a situation where the strongest limitation to understand our surroundings may become, not from the lack of data, or the appropriate technology, but from our own state of mind.