New Zealand relies upon phosphorus (P) to sustain agricultural productivity. However, P loss from farming systems to freshwater ecosystems can promote eutrophication; a global problem. A disproportionate amount of P and sediment is transported from farm systems to freshwaters via ephemeral streams (overland flow) during short of periods of time. Ephemeral streams flow over landscapes (e.g. depressions in paddocks which are usually dry) during intense rainfall events which produce surface runoff. Treating the sheer volume of water leaving a catchment during these events presents a challenge and many mitigation approaches struggle to cope with such large discharges over short periods of time. The objective of this MSc research was to quantify the performance of a new type of detainment bund (DB) being trialled in the Lake Rotorua catchment that was designed to intercept surface runoff by ponding it behind a low profile earth bund (c. 1.5 m high). The aim was to promote settling of suspended sediments and associated particulate P (PP) in the DB basin (onto the pasture). Ponded water was released slowly (via a floating decant structure) until the pond was completely drained; residence time of water was no more than three days to ensure pastoral production in the ponding area was maintained. Three detainment bunds were constructed on private dairy farms within three sub catchments (Waiteti, Hauraki and Awahou) of Lake Rotorua (Bay of Plenty, New Zealand). The DBs have been initiated as part of a collaboration of Bay of Plenty Regional Council, DairyNZ and Rotorua catchment farmers in a wider P mitigation programme known as the ‘Rotorua P-Project’. Sampling was undertaken from March - September 2012 during which eight rainfall events produced ponding in the DBs. Synthetic grass mats and sediment trays were deployed across the ponding area of each DB to capture sediments which were deposited during ponding periods. Grab samples of in– and out– flowing water were collected at various stages during storm events and analysed for total suspended sediments (TSS), particle size distribution, total P (TP), dissolved reactive P (DRP), total nitrogen (TN) and dissolved inorganic nitrogen to determine changes in water quality over the ponding periods. Water level was recorded to calculate the volume of ponded water and estimates of the mass of sediment and P deposited per event were derived. Total P concentrations of up to 1.6 mg L⁻¹ were recorded in ephemeral stream flow. Results showed that there were significant reductions in TSS concentrations throughout the ponding events. The fastest settling rate (73% reduction in TSS over 43 h) occurred when ponded water comprised a large percentage inorganic material, slower settling rates were associated with high % of organic SS. Particulate nutrient concentrations of water leaving the DBs decreased at fastest rates within the first 20 h of ponding (up to 36% of PP and 42% of PN at Awahou DB) when TSS concentrations were high (>100 mg L⁻¹). Similar reductions were observed at the other sites (with lower TSS) but settling rates were typically half this maximum observed rate. The sediment retained by the three DBs in this study was enriched with P (average 2080 mg P (kg dw) ⁻¹) relative to the benthic sediments of Lake Rotorua and the alluvium of the Waiteti Stream. Phosphorus in the deposited sediments was associated with metal cations such as Fe and Mn; this indicates that such PP which is present in redox–sensitive forms is potentially bioavailable in Lake Rotorua during periods of lake stratification which lead to anoxia in the hypolimnion. The largest retention of sediment and P (2,749 kg and 6.08 kg respectively) during the study period occurred when a large rainfall and runoff event in July coincided with the recent complete removal of vegetation in an upstream paddock sown with a winter forage crop. The setting of the Hauraki DB was the most representative of a typical dairy farm and the average deposition per sampled event was 0.261 kg P⁻¹. Using this figure it was estimated, that over a 20 year period, 28 kg of P could be retained (given the same hydrological and catchment characteristics as 2012). This equates to a saving of c. $28,000 if the P was to be removed by in-lake restoration methods. Detainment bunds did not attenuate dissolved nutrients (DRP and DN) during ponding in most cases; however, DRP was inversely correlated with TSS when the TSS concentrations were high. In some ponding events DRP increased, this was likely due to net desorption from suspended sediments. An investigation of the soil P concentrations around the ponding area of an old (12 y) detainment dam built for flood control revealed a significant decrease in Olsen P with distance from the dam wall, indicating that historic ponding had deposited sediment enriched with P in the ponding area (Olsen P ranged from 119 mg L⁻¹ inside to 41 mg L⁻¹ outside of the ponding area). There is a potential for DB ponding areas to be a P source at certain times if DRP is desorbed from P enriched soils to overlying water, however, in the long term the investigation indicated that they are likely to be a P sink and there may be no need for addition of P fertiliser in the ponding area. Adequate water storage capacity was identified as critical to the design of future DBs. Observations during this research showed that storage ratios should be based on a minimum ratio of 120 m³ of water storage per 1 ha of contributing catchment (to the concrete riser). It is important that the floating decant structures used to drain DBs are designed specifically for the volume of each DB to allow ponded water to drain from the DB within the desired time of ponding. Land owners have tolerated three days ponding with no impact on pastoral production. Detainment bunds can play a pivotal role in moderating the hydrological pathways at the catchment scale by prioritising headwater catchments and slowing down water flow to reduce the loss of nutrients and sediment from pastoral farmland during intense rainfall and runoff events. The level of DB implementation within pastoral landscapes will depend on the willingness of landowners to incorporate them into farm systems. A win-win situation is possible where water quality is improved and pastoral production within the ponding areas is maintained.