“…In geostatistics, this is achieved with block kriging, which not only derives a prediction of the spatial average but also quantifies the associated prediction uncertainty (Cressie, 2006;Lark and Lapworth, 2012;Webster and Oliver, 2007). However, block kriging may be challenging in case observations are transformed prior to geostatistical modelling and predictions are back-transformed afterwards (Cressie, 2006;Cressie, 1993;Orton et al, 2015). In such case one often resorts to spatial stochastic simulation, where the spatial average over a region is approximated by generating many simulated region values as spatial averages of simulated point values (Goovaerts, 2001).…”