Explore the vast range of titles available.

River suspended sediment yield estimates from 233 New Zealand catchments are presented and used to calibrate an empirical, raster-type GIS model for predicting suspended sediment yield from any river in New Zealand. The calibration dataset is mostly based on suspended sediment gaugings and flow records, but includes data from lake and fiord bed sedimentation studies. The model relates sediment yield to the spatial integration of the product of a 'driving' factor and a 'supply' factor. The driving factor is P1.7, where P is the local mean annual precipitation extracted from a precipitation grid. The supply factor depends on an erosion terrain classification that spreads erosion potential by slope and lithology, and to some extent by erosion process. With 74 unique terrains defined for New Zealand, coefficients for each terrain were determined by a range of approaches based on the availability of calibration data. Comparison of measured yields with those predicted by a preliminary calibration procedure highlighted several regions with systematic over-or under-prediction. These regions have differences in land cover, glacial history, tectonic regime, and/or climate that are not incorporated into the erosion terrain classification or the driving factor based on mean annual rainfall. Separate coefficients were determined for the dominant erosion terrains in each special region. With these adjustments, the model explained 97% and 96% of the variance in the measured South and North Island (log-transformed) yields, respectively, while the standard errors of the predictions equated to factors of 1.55 and 1.8 for the South and North Islands catchments, respectively. The factorial error in predicted yield decreased as catchment area increased. When totalled over all measured catchments, the predicted yield differed from the measured yield by only 2.3% in the South Island and 6.5% in the North Island. Summing measured yields and model-predicted yields from ungauged catchments, the South Island yield is 91 Mt/y and the North Island yield is 118 Mt/y. Natural lakes intercept 14% of the potential sediment yield of the South Island but only ~1% of that of the North Island. Hydro-lakes intercept ~3% of the South Island yield and -0.3% of the North Island yield. Overall, floodplains and estuaries intercept only a small percentage of the total sediment delivery to the coast (although this may not be the case for particular floodplains or estuaries). The ~209 Mt/y total suspended sediment yield from New Zealand amounts to ~ 1.7% of the global sediment delivery to the oceans, making New Zealand a significant contributor on a unit area basis by virtue of its steep terrain, high rainfall, and tectonic activity. The sediment yield model is available in an easily used form on the Internet (http://wrenz.niwa.co.nz/webmodel), but in its present form it is not suited to assessing the effects of changes in land cover or land use on sediment delivery because this was not explicitly included as a controlling factor.

Data on spatial and temporal patterns of deposited fine sediment (DS) cover in New Zealand rivers are rare and this makes it difficult to assess DS trends and quantify links between sediment and ecological effects. A rapid, semi-quantitative method for characterising DS cover using class-based visual assessment was applied annually for 6 years (2005–10) at 30 sites in the Motueka River catchment, to explore spatial and temporal variation in DS cover and its relationship with sediment load and trout numbers. Deposited fine sediment cover was low at most sites, apart from some small tributaries with erodible granite geology near to the sampling site. Most sites had little change in DS over time. Large floods or forest harvesting caused discernible, but small, increases in DS cover at some sites. Local delivery of sediment and stream morphology were more important drivers of spatial patterns of DS cover than annual sediment load or characteristics of the upstream contributing catchment. For some sites there was a relationship between temporal variation in sediment load, caused by floods or forest harvesting, and DS cover. While trout abundance in the Motueka River varied considerably between 2005 and 2010, there was little relationship with DS cover. Sampling bias, assessed by repeating measurements at five sites on five days, was not significant indicating the method is highly repeatable. However, there was consistent significant bias between results from two different observers implying the visual assessment method requires training to be reliable.

Recent decades have seen advances in sediment flux modelling in New Zealand, including adaptation of models developed overseas, development of some empirical models specifically for New Zealand, and applications to assess the environmental impacts of erosion. This paper documents and summarises the range of sediment flux models and their application. Particular emphasis is given to the degree to which the models represent the wide range of processes occurring in New Zealand, and also to their representation of shallow landslides. It is concluded that the simple empirical models for mean annual flux give limited insight to processes, and so provide only limited guidance for mitigation and intervention measures. At the other extreme, detailed process models are difficult to run and express in terms of parameters, and they still have difficulty in accommodating the range of relevant processes. Modification of an intermediate-complexity model, SedNet, is proposed as a way forward for further model development in New Zealand.

Can we develop land use policy that balances the conflicting views of stakeholders in a catchment while moving toward long term sustainability? Adaptive management provides a strategy for this whereby measures of catchment performance are compared against performance goals in order to progressively improve policy. However, the feedback loop of adaptive management is often slow and irreversible impacts may result before policy has been adapted. In contrast, integrated modelling of future land use policy provides rapid feedback and potentially improves the chance of avoiding unwanted collapse events. Replacing measures of catchment performance with modelled catchment performance has usually required the dynamic linking of many models, both biophysical and socio-economic—and this requires much effort in software development. As an alternative, we propose the use of variable environmental intensity (defined as the ratio of environmental impact over economic output) in a loose coupling of models to provide a sufficient level of integration while avoiding significant effort required for software development. This model construct was applied to the Motueka Catchment of New Zealand where several biophysical (riverine water quantity, sediment, E. coli faecal bacteria, trout numbers, nitrogen transport, marine productivity) models, a socio-economic (gross output, gross margin, job numbers) model, and an agent-based model were linked. An extreme set of land use scenarios (historic, present, and intensive) were applied to this modelling framework. Results suggest that the catchment is presently in a near optimal land use configuration that is unlikely to benefit from further intensification. This would quickly put stress on water quantity (at low flow) and water quality (E. coli). To date, this model evaluation is based on a theoretical test that explores the logical implications of intensification at an unlikely extreme in order to assess the implications of likely growth trajectories from present use. While this has largely been a desktop exercise, it would also be possible to use this framework to model and explore the biophysical and economic impacts of individual or collective catchment visions. We are currently investigating the use of the model in this type of application.

Morphological change in gravelly river channels and active slope environments can be used to derive lower-bound estimates of sediment flux between successive dates on which the morphology of channels or slopes was measured. This paper provides examples of such morphological budgets derived from analysis of digital elevation models built from ground survey data acquired from three sites in the upper Motueka River, and the Tarndale Fan and Gully system in the upper Waipaoa catchment. We demonstrate the utility of such an approach in providing information on processes operating in channel and slope environments at the same time as estimating volumes of sediment flux.

The effects of various groundwater abstraction scenarios on stream flow and groundwater levels in the Upper Moteuka River catchment have been investigated using a groundwater–river interaction model. The model operated on a daily time step and had two components: a FEFLOW groundwater model to simulate groundwater losses and gains from the river, and a custombuilt river model to route river flow and to calculate river water levels using Manning's equation. For the 'base case', the model input included climate and abstraction data for the period 1 July 2001 to 30 June 2003, and the model was calibrated by adjusting hydraulic conductivity and streambed conductance to achieve a good match to observed groundwater levels, river flows and river water levels measured in this same time period. Thereafter the calibrated model was then used to predict groundwater levels, river flows and river water levels for six different scenarios: 1) actual groundwater abstraction at the end of 2008; 2) groundwater abstraction assuming the maximum take specified on each existing water permit; 3) groundwater abstraction assuming additional permitted takes sufficient to allow complete irrigation over all potentially irrigable land; 4) actual groundwater abstraction at the end of 2008 in roughly half of the model area combined with complete irrigation of all potentially irrigable land in the remainder of the model area; and 5) actual groundwater abstraction at the end of 2008 assuming river bed elevations were 0.3 m higher or lower than current elevations. For all scenarios, modelled mean and median river flows during the irrigation season were at most 5% less than for the base case, and differences outside the irrigation season were even smaller. Only the third of the tested scenarios resulted in modelled river flows that may breach minimum flow requirements at some locations. For all scenarios, modelled groundwater levels were no more than 1 m below the base case, with the largest drops occurring at four locations along the margins of the catchment and near pumping centres. The results suggest that a groundwater pumping scheme could incorporate existing and new abstraction while still ensuring that estimated low flow conditions are not breached during the irrigation season, and the scenario modelling results can be used by Tasman District Council for groundwater allocation management.

New Zealand forests burn less frequently than tussock grasslands, heath or shrublands. Species composition, past disturbance and stand condition determine inflammability and fuel load, and consequent fire intensity and spatial extent. Before people arrived, fires were ignited by lightning during drought years on the eastern sides of both islands. Volcanism occurring every 300-600 years was associated with fires in the central North Island. A review of radiocarbon-dated charcoal from the eastern South Island, and of evidence for fire in pollen profiles from the North Island, provide the basis for an assessment of fire frequency. Forest fires have occurred on both New Zealand's islands throughout the Holocene at least every few centuries, until the last millennium when frequency increased. The 'return time' of fire at any one place in the forested landscape was probably one or two millennia. Burned areas usually succeeded to forest again before the next inflagration. Consequently fire adaptation is infrequent in the New Zealand flora, and Polynesian forest clearance was rapid and largely permanent. There is an indication of an increase in fire frequency in the late Holocene, and a clear signal associated with people approx. 700 years BP. Separating the earliest anthropogenic fires from the background level of natural burning will be difficult without additional evidence.

New Zealand treelines have been well studied over the past few decades. Peter Wardle carried out extensive observational and experimental studies of their ecology and suggested that the length and warmth of summer was critical in permitting alpine trees to make sufficient growth to survive winter (Wardle 1985a). He also showed that New Zealand treelines were low when compared with global treelines, in particular with those of southern South America, and that exotic Pinus contorta could grow up to 300 m above the indigenous treeline (Wardle 1985b, 2008). A later global study demonstrated that warmth of the growing season was

Our evaluation of pre-settlement Holocene (10 000–1000 BP) fire, using radiocarbon-dated charcoals and pollen and charcoal spectra in pollen diagrams, concludes that fires were infrequent and patchy in the eastern South Island of New Zealand. Charcoal radiocarbon dates point to three broad phases of fire frequency: infrequent patchy fires from 10 000 to 2600 BP; a slightly increased frequency between 2600 and 1000 BP; and an unprecedented increase of fires after 1000 BP, which peaked between 800 and 500 BP. We suggest that natural fire was driven more by vegetation flammability (with ignitibility and combustibility components) than climate within this rain-shadow region, that plant chemistry principally determined fire frequency, and that topography determined the extent of fire. The review suggests that there were rare spatial and temporal instances of a feedback relationship between fire and early-successional grasses in eastern South Island. This occurred only within narrow-range, cool environments, whose equilibrium communities were of flammable, phenolic-rich woody species and grasses, and was predominantly in the late pre-settlement period. Elsewhere, grasses and herbs were understorey components to otherwise low-flammability, hardwood forest and scrub.

This impressive collection celebrates the work of Peter Kershaw, a key figure in the field of Australian palaeoenvironmental reconstruction. Over almost half a century his research helped reconceptualize ecology in Australia, creating a detailed understanding of environmental change in the Late Pleistocene and Holocene. Within a biogeographic framework one of his exceptional contributions was to explore the ways that Aboriginal people may have modified the landscape through the effects of anthropogenic burning. These ideas have had significant impacts on thinking within the fields of geomorphology, biogeography, archaeology, anthropology and history. Papers presented here continue to explore the dynamism of landscape change in Australia and the contribution of humans to those transformations. The volume is structured in two sections. The first examines evidence for human engagement with landscape, focusing on Australia and Papua New Guinea but also dealing with the human/environmental histories of Europe and Asia. The second section contains papers that examine palaeoecology and present some of the latest research into environmental change in Australia and New Zealand. Individually these papers, written by many of Australia's prominent researchers in these fields, are significant contributions to our knowledge of Quaternary landscapes and human land use. But Peopled Landscapes also signifies the disciplinary entanglement that is archaeological and biogeographic research in this region, with archaeologists and environmental scientists contributing to both studies of human land use and palaeoecology. Peopled Landscapes reveals the interdisciplinary richness of Quaternary research in the Australasian region as well as the complexity and richness of the entangled environmental and human pasts of these lands. - Prof. Peter Hiscock, The Australian National University