Resource Estimation, Astrophotography and MREC 2023

One the night before MREC 2023 I thought I’d gather my thoughts and prepare. You see it’s not everyday we get a group of people together to discuss the arcane art and science of resource estimation. And that might be a good thing - the world can only cope with that sort of specialisation in small doses!

Rather than talk directly about the conference I would like to draw some analogies. Now, arguing by analogy can be a dangerous approach and it’s subject to all kinds of confusion, but at times it can be useful. So let’s begin.

What is resource estimation? I called it an art and science in the opening paragraph. We draw together a multitude of threads pulling on our experience, expertise, knowledge to predict properties of the earth’s crust. We forecast the location and quantity and quality of elements with varying degrees of accuracy and diligence. These forecasts, predictions or estimates are grist for the mill, used and abused by an unpredictable set of stakeholders for an unpredictable period of time. You never know when an estimate bearing your name is going to pop back up again.

That’s kind of weird and partially disturbing. Why? Because an estimate is more than the digital essence of the model, the report the public statement. An estimate is a snapshot. A moment frozen in time. It encapsulates the beliefs and paradigms of the day as much as it rests on the ability of the practitioner. 

Let me draw that fraught analogy. Few people know of one of my hobbies. I like capturing images of deep-space objects (DSO). It’s a type of astrophotography. Now, believe it or not there are a lot of similarities between DSO imaging and resource estimation. Both are based on sparse data. Both have high degrees of uncertainty. Both are affected by noise, variation and bias. 

And both are very dependent on the skills of the practitioner and the equipment in their arsenal. 

The type of DSO I make images of are not visible to the human eye. They depend on the wonder of modern digital sensors in specialised cameras. These sensors watch ancient photons from thousands to millions of years ago and their digital footprint is left imprinted along with an array of almost mind numbing noise. You have the noise from the sensor itself. Noise that can be divided into :

1. a random pattern (or bias) that occurs when the sensor takes any reading. The act of reading the data creates noise! 
2. thermal noise that builds up from the heat generated during the long exposures while the sensor is capturing those infrequent photons
3. noise in the optical chain. Things like dust on lenses and sensors, errors in the lens diffraction or other aberrations. 

Over the year professional and amateur astrophotographers have learned to manage these noise inputs to enhance the signal-to-noise ratio. Have we done the same in resource estimation? Arguably no. If anything we tend to ignore or downplay the impact of various sources of noise, believing that if we ignore it, it might go away.

The next part of DSO imaging that is similar to resource estimation is the image manipulation itself. When those wonderful sensors capture those sparse photons and create their digital magic our human eyes still cannot see anything particularly spectacular. Even over many hours (my longest ‘integration time’ is ~18 hrs) the data is bland, uninformative and unlikely to appear of any value. The stars are probably over exposed (blown) and everything else looks a boring black, or worse, a dull grey/green due to light pollution. The art and science of DSO imaging is in taking that unlikely data and modelling it. The wavelengths or the captured light are all tightly compressed and the astrophotographer must manipulate the frequency distribution to ’stretch’ the histogram, slowly pulling detail from the data with a subtle hand, magnifying the contrast and revealing the unseen detail. 

That is not too unlike a resource estimator who takes the spatially sparse data and analyses its nature, looking for patterns, similarities and differences. Taking those observations and building models (which inevitably turn into digital images of one form or another) with care to not over-exaggerate or under-exaggerate the variation in the rocks nor to introduce a bias that may cast an unwelcome hue over the estimate.

This DSO image manipulation is tricky and very dependent on the skill of the operator. I know. I’m still on a very steep part of the learning curve. But here’s one thing that is different between DSO imaging and resource estimation. On cloudy nights, when the sky is obscured, the DSO astrophotographer will go back to their old data and reprocess it. Taking a different path to show new detail, new structure, new insight. Like the images below - those are all from the exact same set of data. All a series of images collected on the same night. The difference is in the ‘post-processing’ or modelling. My early images are on the left and later images on the right. You can see the change as practice and experience kick in. 

You can also see how different aspects are highlighted. In some the stars are more dominant. In other the nebula are the hero. That too, is like resource estimation. You see, when I post those images where others involved in this hobby can comment their judgements of the image quality are varied. Some prefer one ‘style’ others a different emphasis. And… those judges are not wrong! Let’s face it. None of us know exactly what these structures are really like or how they would look if we could directly observe them. They are all beyond the realm of or direct observation and we only know them through the images models we create - our estimates if you like. And that’s just like a resource estimate. It too is something we can never know with certainty. We can never measure the resource directly as we are always dependent on some sort of extraction process. Thus, at best, we measure the combination of the resource and the processes used during extraction.

Yes, there are parallels here. A different field of endeavour that faces many similar challenges. It has different solutions too and maybe we can learn from that. Sometimes you need to let go of the old paradigm and open yourself to new ideas, new approaches and new practices. Sometimes you also need to admit you don’t know what you don’t know. Sometimes you need to stare reality right in the eye, take a deep breath and let all those preconceived ideas wash away in a flood of uncertainty.

Hope I see you at the conference. If not, I hope you can take these words and see their value. After all, we do ourselves an injustice if we think we have conquered the complexity of modelling the earth’s precious resources from what amounts to almost no directly measured data.