Alteration envelopes around volcanogenic massive sulphide (VMS) deposits are commonly several kilometres larger than the associated mineral deposit, making them suitable exploration targets. Furthermore, these alteration envelopes are usually zoned with different alteration types and variable alteration intensity at different distance to mineralisation, whereby they can be used as ore vectors. However, alteration mapping is subjective and can vary between geologists. Therefore, quantitative approaches using whole rock lithogeochemistry are useful for mapping alteration. One such approach uses mass balance calculations to quantify mass changes in mobile elements resulting from alteration. A challenge with this technique is that it relies on sampling least-altered volcanic rocks representative of precursor compositions. Crucially, mass balance calculation depends on how well constrained the least-altered samples are and as such, these samples are chosen with great care; however, in some exploration projects sub-optimal least-altered sample choices are made. Furthermore, uncertainties stemming from sampling- and analytical errors and uncertainty in modelling fractionation curves for major mobile elements influences the certainty of variables going into mass balance calculations. It is therefore of interest to study the impact of least-altered sample choices and error propagation on the final mass balance calculation allowing future studies to make more informed decisions regarding selection criteria for least-altered samples.

This study uses the Rävliden North deposit in the western part of the Palaeoproterozoic Skellefte district, Northern Sweden, as a natural laboratory to test the impact of least-altered sample choices. This deposit is a relatively recent Zn-Pb-Ag-Cu VMS discovery approximately 4 km west of the currently operated Kristineberg mine. Alteration envelopes of with varied amounts of quartz, sericite, chlorite and talc are commonly associated with VMS deposits in the Skellefte district, and with an alteration intensity locally strong enough to eradicate textures of the original lithofacies. Furthermore, the deposits are modified by polyphase deformation, greenschist to amphibolite facies metamorphism, and remobilization. Combined these make stratigraphic analysis and lithofacies mapping difficult, which motivates the use of lithogeochemical techniques when exploring for these deposits. In the Rävliden North area, two styles of VMS mineralization types occur: 1) a semi-massive to massive Sp+Pyh+Gn±Py hosted in the Rävliden formation in upper parts of the Skellefte group (SG), and 2) a stringer Ccp+Pyh+Py mineralisation occurring in both the SG and Rävliden formation.

By the comparison of mass change results from calculations using two datasets with differently constrained least-altered sample choices this study concludes that absolute mass changes are sensitive to different least-altered sample choices and that with a 50% confidence interval on regression an uncertainty in mass change of approximately 0.5 wt.% for MgO, FeO, CaO, and 0.2 for K2O and Na2O, and 5 wt.% for SiO2 is found. However, regardless of the least-altered sample choice, qualitative recognition of ore vectors is possible. Furthermore, this study find the following ore vectors to Rävliden North: 1) semi-regional Na2O and CaO loss, 2) distal K2O gain, 3) proximal K2O loss, 4) proximal CaO gain is associated with the Sp+Pyh+Gn±Py mineralisation, 5) proximal MgO and Fe2O3 gains associated with both the Sp+Pyh+Gn±Py and Ccp+Pyh+Py mineralisation, and 6) proximal erratic gains and losses in SiO2 is associated with both mineralisation types.