Adverse effects from sediment and heavy metals have been observed in the Heathcote catchment, which is diverse in its land use activities. Stormwater management improvements are planned for the catchment through the Heathcote Stormwater Management Plan. Contaminant load monitoring and modelling for subcatchments in the Heathcote were undertaken to help inform the stormwater management policies and planning. The UC’s event-based contaminant load model, MEDUSA (Modelled Estimates of Discharges for Urban Stormwater Assessments), that predicts the amount of total suspended solids (TSS), and total and dissolved copper and zinc generated by individual roof, road and carpark surfaces, was employed for the modelling. Stormwater monitoring of key impermeable surfaces was used to calibrate the model and also quantify the chemical speciation of the contaminants (i.e. particulate or dissolved form), important for assessing appropriate future treatment or mitigation strategies. Stormwater runoff monitoring and predictive modelling (using MEDUSA) was previously conducted in the Okeover and Addington subcatchments of the Avon Catchment. In those studies, four roof types, three road types and three carpark classifications were monitored and modelled. In this study, stormwater runoff quality was monitored from eight different impermeable surfaces in the Heathcote catchment over 9 rainfall events from July to November 2016. These sites represented typical surfaces in the catchment: a new Coloursteel® roof, an older Coloursteel® roof, a concrete roof, a galvanized painted roof, three roads (local, collector, minor arterial) and a commercial/light industrial carpark. First flush (first 1 L of runoff) and steady state samples were analysed for TSS and total and dissolved zinc and copper. Data from the monitoring campaign was analyzed and then used to refine MEDUSA to Heathcote conditions. The model was applied to estimate zinc loads from roofs for the Heathcote catchment as a whole, as well as from six individual subcatchments representing industrial areas (Curries and Jardens Drains; Awatea), mixed use areas (Curletts Drain; Waltham) and mostly residential areas (Jacksons Creek; Wilderness Drain). Predictions of contaminant loads were obtained for each rainfall event sampled in 2016. Additionally, predictive simulations were conducted for all events for years between 2011-2016 to ascertain differences as a function of variable weather conditions. The galvanized roof surfaces produced significantly more zinc than other surfaces. Coloursteel® Old and Galvanised Painted first flush runoff contributed some of the highest zinc concentrations measured in recent Christchurch untreated stormwater sampling. First flush concentrations from the new Coloursteel® roof were consistently lower than the steady state concentrations from the old Coloursteel® roof. Similarly, zinc concentrations from the galvanised painted roof were higher than the new Coloursteel® roof, but lower than the old Coloursteel® roof. The data also clearly show that the majority of zinc from the four roof types is in the dissolved form, substantiating previous monitored data in Christchurch. These data confirm that the key mechanism for zinc generation from roofs is direct dissolution of the roof material, enhanced and sustained by the exposure and breakdown of the galvanizing layer through weathering. Zinc measured in concrete roof runoff is believed to originate from galvanised components in the guttering and downpipes rather than from atmospheric deposition alone. Therefore, while concrete and other non-metallic roofs may not contribute large zinc loads to stormwater runoff, some zinc is dissolved from their galvanised drainage components, which may be something to consider in management decisions about roof replacements along with roof condition. Because zinc was defined as the focus of the study, total zinc loads were predicted using MEDUSA. Modelling results revealed that there is a clear difference in the rate at which total zinc is derived from each roof type, with concrete and Coloursteel® roofs yielding the least amount of zinc (per area) in roof runoff. Zincalume® and painted Galvanised roofs released more than double the amount (per area) of concrete and Coloursteel® roofs, but not as much as unpainted galvanised roofs. The data highlight the availability of zinc from roofs (with metallic surfaces) to stormwater runoff and the positive effect of painting these surfaces to immobilize some of the zinc. The yearly scenario results reveal the influence of variable wet weather conditions (including rainfall pH, antecedent dry days, rainfall intensity and duration) on zinc runoff from roofs. Despite the relatively low proportion (7 %) of roofs within the Heathcote Catchment that are defined as poorly painted or unpainted, they consistently contribute more than 30 % of the total zinc load from roofs in each year. Waltham (mixed landuse) roofs, which make up 29 % of the catchment and comprise the highest proportion (16 ha) of unpainted galvanized roofs, contribute between 2.2 and 7.6 net kg TZn/event to stormwater runoff. Similarly, Wilderness Drain (residential landuse) roofs, which make up 26 % of the catchment and comprise 12 ha of unpainted galvanized roofs, 34 ha of painted galvanized roofs and 27 ha of Coloursteel® roofs, produces nearly the same net zinc loads (2.0 -7.9 TZn kg) per rain event as Waltham. These disaggregated data are important because they highlight that the proportional area of specific roof types (e.g. unpainted galvanized) is a clear determinant of how much total zinc can be expected in roof runoff rather than assuming greater contributions from a more industrial/commercial area alone. Furthermore, depending on the condition of that roof material, a range of lower or higher zinc loads can be expected from roof runoff during rain events. Changes (as modelled scenarios) in proportional roof areas from the current status would result in significant reductions of total zinc runoff from roofs in the Heathcote subcatchments across all the modelled years, with some variability between years due to the influence of rainfall parameters. This reduction is more pronounced at the higher ranges for each scenario. A change in proportional zinc loads in different subcatchments results from the change in their proportional areas (and condition), highlighting the value in examining specific subcatchment responses to variable modelling scenarios.