Hi @Poodleface, glad it was of interest. Ha ‘dreadnought themed...

Hi @Poodleface, glad it was of interest. Ha ‘dreadnought themed budgie smugglers’ coming soon to the merch store.

Regarding the different levels of processing – to be honest it’s impossible to know. To be fair they had already processed to what they call level 3, but with no further interpretation. This is enough to provide individual and multi-mineral maps along with rule class maps.
The additional level of processing adds further mineral interpretation, including compositional mapping. The additional compositional mapping is an important step particularly with phyllosilicate minerals as it allows you to identify potential variation in mineral chemistry that can be related to temperature and pH due to hydrothermal fluid flow. But the biggest change in the original processing vs 2015 is that they asked HyVista to pull it all together and interpreted it for them.
Why they didn’t get that done originally – who knows. It depends where exploration sat in the priorities list at the time, budgets, or simply knowing what questions to ask of the data. Maybe they assumed they had in-house capacity to do it (or at least wanted to try) and it was more involved than initially thought. In many cases at the exploration stage, the low hanging fruit is sought. Reaching higher can be a fruitless exercise.

We can see similar examples of data under-utilisation everywhere, companies flying extensive airborne magnetic surveys only to use a couple of RTP/1vd derivatives. That might be all they need/want at the time. Another company will come along and re-process the data, use different filters and generate different derivatives or take it to the next level and incorporate geological info to produce a seem-less 3D or even 4D lithostratigraphic model. Same goes for geochemical data, collection of high-quality 4-acid digest data that might end up sitting in databases barely looked at vs full geochemical classification of the rocks and building that into a working geology model.
I guess my point being, there is a wide range of application with spectral data and it comes down to knowing what is possible (and the limitations).

In terms of the Neodymium oxide mineral map:
The overarching statement that I would put upfront is, we shouldn’t rely on any single spectral output in isolation, multiple products need to be used and then considered along with geophysics and geology, geomorphology etc.

So generally speaking, the presence or more importantly the absence of an anomaly in the Neodymium oxide spectral map is far from definitive…imo
It’s basically trying to fingerprint the diagnostic absorption features associated with monazite and neodymium oxide in the VNIR range using spectra collected at much better resolution (lab-based instruments) then attempting to isolate those features in the HyMap data…..HyVista made note of this and the concerns:
“Though there are absorption features in the spectral range where the USGS spectral library spectrum of neodymium has distinct features there is not a clear match between these spectra. The distribution of the unit mapped by this spectrum is widespread and not consistent with a dyke source; it requires field checking”
This is to be expected, even in higher-grade ironstones the abundance of monazite is low…relative to the other minerals. So, this is why it wasn’t used as one of the primary inputs for identifying ironstones.
I chose to make a note of it - only because it was interesting at the C3-C5 area.The lack of Nd-spectral signal at C3 could be the result of what has shown (thus far) to be a lower grade REE. So, when that is weathered and further diluted at surface it is simply too low abundance to be recognised spectrally. Or it might be due to the influence of transported alluvium being so close to the river (figure below).


Similarly, the discrete anomaly towards the south (C5 area) is located further away from the river, so possibly less transported dilution. It might just be picking up a localised weathering effect over the residual sub-crop (making it look anomalous relative to surrounds) or it might be primary enrichment associated with mineralisation…Again, it’s difficult to tell especially without ground-truthing.

To me the benefit of this hyperspectral data isn't necessarily in the direct identification (i.e. neodymium mineral map etc) but when its used in combination with the geophysics and mapping to pull apart the geology and then look for subtle structural or hydrothermal alteration features as a way of prioritising drill targets...this is where I see the most vale at the regional scale.