Few studies currently exist that aim to validate a proxy chironomid-temperature reconstruction with instrumental temperature measurements. We used a reconstruction from a chironomid percentage abundance data set to produce quantitative summer temperature estimates since AD 1650 for NW Iceland through a transfer function approach, and validated the record against instrumental temperature measurements from Stykkishólmur in western Iceland. The core was dated through Pb-210, Cs-137 and tephra analyses (Hekla 1693) which produced a well-constrained dating model across the whole study period. Little catchment disturbance, as shown through geochemical (Itrax) and loss-on-ignition data, throughout the period further reinforce the premise that the chironomids were responding to temperature and not other catchment or within-lake variables. Particularly cold phases were identified between AD 1683–1710, AD 1765–1780 and AD 1890–1917, with relative drops in summer temperatures in the order of 1.5–2°C. The timing of these cold phases agree well with other evidence of cooler temperatures, notably increased extent of Little Ice Age (LIA) glaciers. Our evidence suggests that the magnitude of summer temperature cooling (1.5–2°C) was enough to force LIA Icelandic glaciers into their maximum Holocene extent, which is in accordance with previous modelling experiments for an Icelandic ice cap (Langjökull).