The magnocellular vasopressin neurons of the supraoptic nucleus and the paraventricular nucleus of the hypothalamus project axonal terminals to the posterior pituitary, where they secrete vasopressin hormone into the blood, acting at the kidneys to reduce water loss, as part of the homeostatic system regulating osmotic pressure. The neurons respond to synaptic input and depolarising currents generated by osmotic pressure, with increasing activity generating a distinctive phasic firing pattern, consisting of long bursts and silences lasting tens of seconds. We have previously modelled this phasic spike firing mechanism, and have now integrated our spike model with a model of the spike triggered calcium driven secretion mechanism. The secretion mechanism is highly non-linear, showing both frequency facilitation, and fatigue. The secretion response is reduced after about 20s of intense stimulation, as the readily releasable pool of vasopressin vesicles is depleted, and can be reversed by 20-30s of quiescence. Our secretion model uses a simple representation of the calcium dynamics driving vasopressin release, synthesis, and transport to reproduce these effects and fit in vitro secretion data. We suggest that vesicles synthesised in the cell body are transported to a reserve pool, before being transported to the readily releasable pool docked at the cell membrane and use this to explain the observed dynamics. Combined with the spike model we have begun to investigate the relation between phasic firing and the secretion response, studying both individual neurons, and as a population. Conventionally the purpose of phasic firing is to optimise to secretion response per spike, by mixing intense stimulation with silent periods allowing recovery. We reproduce this effect with the model but also show that it gives no great advantage over non-phasic cells, and suggest that the sequence of pools is more important to buffer and maintain response while the central vasopressin store is being depleted. We also show that acting within an asynchronous population, phasic firing serves to give a much more linear response to increasing osmotic input compared to non-phasic model cells, matching the linear relationship observed in vivo.