Simulating wetland impacts on stream flow in southern Africa using a monthly hydrological model

The processes that occur in wetlands and natural lakes are often overlooked and not fully incorporated in the conceptual development of many hydrological models of basin runoff. These processes can exert a considerable influence on downstream flow regimes and are critical in understanding the general patterns of runoff generation at the basin scale. This is certainly the case for many river basins of southern Africa which contain large wetlands and natural lakes and for which downstream flow regimes are altered through attenuation, storage and slow release processes that occur within the water bodies. Initial hydrological modelling studies conducted in some of these areas identified the need to explicitly account for wetland storage processes in the conceptual development of models. This study presents an attempt to incorporate wetland processes into an existing hydrological model, with the aim of reducing model structural uncertainties and improving model simulations where the impacts of wetlands or natural lakes on stream flow are evident. The approach is based on relatively flexible functions that account for the input–storage–output relationships between the river channel and the wetland. The simulation results suggest that incorporating lake and wetland storage processes into modelling can provide improved representation (the right results for the right reason) of the hydrological behaviour of some large river basins, as well as reducing some of the uncertainties in the quantification of the original model parameters used for generating the basin runoff. Copyright © 2013 John Wiley & Sons, Ltd.     

On the prediction of the Toce alpine basin floods with distributed hydrologic models

With the objective of improving flood predictions, in recent years sophisticated continuous hydrologic models that include complex land‐surface sub‐models have been developed. This has produced a significant increase in parameterization; consequently, applications of distributed models to ungauged basins lacking specific data from field campaigns may become redundant. The objective of this paper is to produce a parsimonious and robust distributed hydrologic model for flood predictions in Italian alpine basins. Application is made to the Toce basin (area 1534 km²). The Toce basin was a case study of the RAPHAEL European Union research project, during which a comprehensive set of hydrologic, meteorological and physiographic data were collected, including the hydrologic analysis of the 1996–1997 period. Two major floods occurred during this period. We compare the FEST04 event model (which computes rainfall abstraction and antecedent soil moisture conditions through the simple Soil Conservation Service curve number method) and two continuous hydrologic models, SDM and TDM (which differ in soil water balance scheme, and base flow and runoff generation computations). The simple FEST04 event model demonstrated good performance in the prediction of the 1997 flood, but shows limits in the prediction of the long and moderate 1996 flood. More robust predictions are obtained with the parsimonious SDM continuous hydrologic model, which uses a simple one‐layer soil water balance model and an infiltration excess mechanism for runoff generation, and demonstrates good performance in both long‐term runoff modelling and flood predictions. Instead, the use of a more sophisticated continuous hydrologic model, the TDM, that simulates soil moisture dynamics in two layers of soil, and computes runoff and base flow using some TOPMODEL concepts, does not seem to be advantageous for this alpine basin. Copyright © 2006 John Wiley & Sons, Ltd.     

Effect of variations in net basin supply on lake levels and outflows

Starting from a simplified version of the water balance equation, it is shown how the levels and outflows from a open lake can be related to the magnitude and statistical characteristics of the annual net basin supply series, which is defined as the net inflow arising from tributary inflows, rainfall on the lake surface and the lake evaporation. Using the models, estimates are derived for the response time for several large open lakes in Africa and Asia. Also, the observed net basin supply series for three major lakes (Victoria, Malawi and Toba) are investigated in detail. It is suggested that, on an annual basis, the assumption of a random normally distributed series of net basin supply can provide a reasonable first guess for the equilibrium response of a lake and for the likely range of variation of levels. These estimates are likely to be useful in water resources assessments and for investigations of the effects of land use and climate change on lake levels. The effects of departures from this simplified approach are discussed and are illustrated with some practical examples. © 1998 John Wiley & Sons, Ltd.     

Lake metabolism scales with lake morphometry and catchment conditions

We used a comparative data set for 25 lakes in Denmark sampled during summer to explore the influence of lake morphometry, catchment conditions, light availability and nutrient input on lake metabolism. We found that (1) gross primary production (GPP) and community respiration (R) decline with lake area, water depth and drainage ratio, and increase with algal biomass (Chl), dissolved organic carbon (DOC) and total phosphorus (TP); (2) all lakes, especially small with less incident light, and forest lakes with high DOC, have negative net ecosystem production (NEP < 0); (3) daily variability of GPP decreases with lake area and water depth as a consequence of lower input of nutrients and organic matter per unit water volume; (4) the influence of benthic processes on free water metabolic measures declines with increasing lake size; and (5) with increasing lake size, lake metabolism decreases significantly per unit water volume, while depth integrated areal rates remain more constant due to a combination of increased light and nutrient limitation. Overall, these meta-parameters have as many significant but usually weaker relationships to whole-lake and benthic metabolism as have TP, Chl and DOC that are directly linked to photosynthesis and respiration. Combining water depth and Chl to predict GPP, and water depth and DOC to predict R, lead to stronger multiple regression models accounting for 57–63% of the variability of metabolism among the 25 lakes. It is therefore important to consider differences in lake morphometry and catchment conditions when comparing metabolic responses of lakes to human impacts.     

Lateral and subsurface flows impact arctic coastal plain lake water budgets

Arctic thaw lakes are an important source of water for aquatic ecosystems, wildlife, and humans. Many recent studies have observed changes in Arctic surface waters related to climate warming and permafrost thaw; however, explaining the trends and predicting future responses to warming is difficult without a stronger fundamental understanding of Arctic lake water budgets. By measuring and simulating surface and subsurface hydrologic fluxes, this work quantified the water budgets of three lakes with varying levels of seasonal drainage, and tested the hypothesis that lateral and subsurface flows are a major component of the post‐snowmelt water budgets. A water budget focused only on post‐snowmelt surface water fluxes (stream discharge, precipitation, and evaporation) could not close the budget for two of three lakes, even when uncertainty in input parameters was rigorously considered using a Monte Carlo approach. The water budgets indicated large, positive residuals, consistent with up to 70% of mid‐summer inflows entering lakes from lateral fluxes. Lateral inflows and outflows were simulated based on three processes; supra‐permafrost subsurface inflows from basin‐edge polygonal ground, and exchange between seasonally drained lakes and their drained margins through runoff and evapotranspiration. Measurements and simulations indicate that rapid subsurface flow through highly conductive flowpaths in the polygonal ground can explain the majority of the inflow. Drained lakes were hydrologically connected to marshy areas on the lake margins, receiving water from runoff following precipitation and losing up to 38% of lake efflux to drained margin evapotranspiration. Lateral fluxes can be a major part of Arctic thaw lake water budgets and a major control on summertime lake water levels. Incorporating these dynamics into models will improve our ability to predict lake volume changes, solute fluxes, and habitat availability in the changing Arctic. Published 2016. This article is a U.S. Government work and is in the public domain in the USA.     

21世纪以来泛北极湖泊水位变化时空特征及原因探讨

在全球气候变暖和极端气候事件增加的背景下,流域水文循环过程受到的影响越来越强烈,导致湖泊水位变化表现出复杂的时空特征。而泛北极地区是地球上湖泊数量与面积分布最为集中的区域之一,该地区湖泊对气候变化响应非常敏感。因此,了解这些湖泊近期水文变化特征十分必要。本研究共搜集了36个泛北极大型湖泊(>500 km²)基于遥感或站点观测的近20年水位数据,分析其时空变化特征。本文使用线性回归模型来估算湖泊水位的变化趋势,进而利用皮尔逊相关分析了其主要水文影响变量和大气环流机制,并运用Mann-Kendall突变检验法探讨了水位突变的原因。结果表明,泛北极湖泊的水位整体上呈现不同程度上升(平均速率为0.013 m/a),有23个(64%)湖泊的水位呈上升趋势;研究湖泊中有10个通过90%统计显著性检验。其中,水位上升速率最大的湖泊是位于哈萨克斯坦的腾吉兹湖,上升速率为0.078 m/a。泛北极湖泊水位的波动主要与径流有关,有19个(53%)湖泊的水位波动与径流的增加更为相关;相比而言,位于亚洲的极地湖泊水位的上升与流域蒸发的降低显著相关,尤其是库苏古尔湖。从区域大气环流影响来看,泛北极湖泊水位变化主要与厄尔尼诺-南方涛动有关,其次是北极涛动和北大西洋涛动。本研究有助于加深对泛北极湖泊近20年水位变化规律及气候影响特征的科学理解。     

Tracing contributions to hydro‐isotopic differences between two adjacent lakes in the southern Tibetan Plateau

Hydrological processes of lakes in the Tibetan Plateau are an important indicator of climate change. Due to the high elevation, inaccessibility and limited availability of historical observations, water budget evaluation of typical lake basins has been inadequate. In this study, stable isotopes are used to trace the multiple water sources contributing to two adjacent lakes on the north slope of the Himalayas, Gongmo‐tso and Drem‐tso. The two lakes have nearly the same elevation, lake area and climatic condition. However, the isotopic composition of the two lakes presents significant differences. Qualitative observations attribute the differences to hydrological discrepancies: Gongmo‐tso is a through‐flow lake, whereas Drem‐tso is a terminal lake. Quantitative analyses, including water and isotope mass balance modelling, clarify the fluxes and isotopic compositions among the various hydrological elements. The isotopic composition of input water, calculated as the summation of rainfall and upstream runoff, is estimated using the local meteoric water line (LMWL) combined with the time series of lake water isotope values. The isotopic composition of evaporation is calculated with a linear resistance model using local meteorological data. The results show that Drem‐tso is a closed lake in a hydrological steady state with relatively more enriched lake water isotope values resulting mainly from evaporation. In contrast, through‐flow accounts for more than 88% of the water input into Gongmo‐tso. The large amount of upstream runoff with lower isotopic composition and enrichment due to evaporation are the major contributions to the observed lake water isotope values. Isotopic modelling of the two neighbouring lakes is effective for isotopic and hydrological research in this region with limited in situ observations. Copyright © 2013 John Wiley & Sons, Ltd.     

Prediction of river water temperature: a comparison between a new family of hybrid models and statistical approaches

River water temperature is a key physical variable controlling several chemical, biological and ecological processes. Its reliable prediction is a main issue in many environmental applications, which however is hampered by data scarcity, when using data‐demanding deterministic models, and modelling limitations, when using simpler statistical models. In this work we test a suite of models belonging to air2stream family, which are characterized by a hybrid formulation that combines a physical derivation of the key equation with a stochastic calibration of parameters. The air2stream models rely solely on air temperature and streamflow, and are of similar complexity as standard statistical models. The performances of the different versions of air2stream in predicting river water temperature are compared with those of the most common statistical models typically used in the literature. To this aim, a dataset of 38 Swiss rivers is used, which includes rivers classified into four different categories according to their hydrological characteristics: low‐land natural rivers, lake outlets, snow‐fed rivers and regulated rivers. The results of the analysis provide practical indications regarding the type of model that is most suitable to simulate river water temperature across different time scales (from daily to seasonal) and for different hydrological regimes. A model intercomparison exercise suggests that the family of air2stream hybrid models generally outperforms statistical models, while cross‐validation conducted over a 30‐year period indicates that they can be suitably adopted for long‐term analyses. Copyright © 2016 John Wiley & Sons, Ltd.     

基于Copula函数的连河通江湖泊防洪安全设计:以洪泽湖为例

湖泊作为一种蓄水单元,尤其是大型过水性湖泊,是一种典型的平原型水库,在功能上与山谷型水库具有许多相似之处,但由于其特殊的地理地形构造,使得入湖洪水过程与入库洪水过程存在着较大的差异.在防洪安全设计研究中,山谷型水库关注的多是坝址洪水,即总的入库洪水过程,而对于湖泊来说,还需要关注各个分区的入湖洪水过程对湖区洪水演进的影响.针对大型过水性湖泊入湖洪水特征,本文采用Copula函数构造了多个联合分布函数,提出了一套基于总的入湖洪水过程推导各个分区入湖洪水过程置信区间的方法.以洪泽湖为应用实例,结果表明:1)在联合重现期已知的情况下,该方法能够确定总入湖洪量与洪峰的95%置信区间;2)该方法通过径流相关性分析对入湖河道合并聚类,形成分区入湖过程,既考虑了河道间天然的水文、水力联系,又避免了联合分布函数维度过高的问题;3)在总入湖洪量已知的情况下,该方法能够确定各分区入湖洪量分配95%置信区域.该方法具有较强的统计理论基础,拓展了多变量洪水频率分析技术在水利工程实际中的应用范围.     

HYDRO_GEN: A spatially distributed random field generator for correlated properties

This paper describes a new method for generating spatially-correlated random fields. Such fields are often encountered in hydrology and hydrogeology and in the earth sciences. The method is based on two observations: (i) spatially distributed attributes usually display a stationary correlation structure, and (ii) the screening effect of measurements leads to the sufficiency of a small search neighborhood when it comes to projecting measurements and data in space. The algorithm which was developed based on these principles is called HYDRO_GEN, and its features and properties are discussed in depth. HYDRO_GEN is found to be accurate and extremely fast. It is also versatile: it can simulate fields of different nature, starting from weakly stationary fields with a prescribed covariance and ending with fractal fields. The simulated fields can display statistical isotropy or anisotropy.     

Linear and nonlinear input/output models for karstic springflow and flood prediction at different time scales

Karstic formations function as three-dimensional (3D) hydrological basins, with both surface and subsurface flows through fissures, natural conduits, underground streams and reservoirs. The main characteristic of karstic formations is their significant 3D physical heterogeneity at all scales, from fine fissuration to large holes and conduits. This leads to dynamic and temporal variability, e.g. highly variable flow rates, due to several concurrent flow regimes with several distinct response times. The temporal hydrologic response of karstic basins is studied here from an input/output, systems analysis viewpoint. The hydraulic behaviour of the basins is approached via the relationship between hydrometeorological inputs and outputs. These processes are represented and modeled as random, self-correlated and cross-correlated, stationary time processes. More precisely, for each site-specific case presented here, the input process is the total rainfall on the basin and the output process is the discharge rate at the outlet of the basin (karstic spring). In the absence of other data, these time processes embody all the available information concerning a given karstic basin. In this paper, we first present a brief discussion of the physical structure of karstic systems. Then, we formulate linear and nonlinear models, i.e. functional relations between rainfall and runoff, and methods for identifying the kernel and coefficients of the functionals (deterministic vs. statistical; error minimisation vs. polynomial projection). These are based mostly on Volterra first order (linear) or second order (nonlinear) convolution. In addition, a new nonlinear threshold model is developed, based on the frequency distribution of interannual mean daily runoff. Finally, the different models and identification methods are applied to two karstic watersheds in the french Pyrénées mountains, using long sequences of rainfall and spring outflow data at two different sampling rates (daily and semi-hourly). The accuracy of nonlinear and linear rainfall-runoff models is tested at three time scales: long interannual scale (20 years of daily data), medium or seasonal scale (3 months of semi-hourly data), and short scale or “flood scale” (2 days of semi-hourly data). The model predictions are analysed in terms of global statistical accuracy and in terms of accuracy with respect to high flow events (floods).     

湖泊水质遥感的几个关键问题

我国目前约有面积大于1km2的湖泊有3000个,绝大部分属高叶绿素和高悬浮物浓度水体,属于典型的Ⅱ类水体,物质组成多样,水体的光学辐射传输复杂,且有大范围的光学浅水.我国的湖泊水质/水色遥感虽然取得了一定进展,但借鉴海洋水色遥感的相关理论和经验,还需要解决以下四个关键问题:1)兼顾海洋沿海水质遥感,发展专用的静止卫星湖泊水质遥感器;2)在当前多光谱遥感资料基础上研发高光谱湖泊水质因子提取的遥感定量化模型,提高反演精度;3)深化湖底底质对湖泊水质/水质遥感影响研究,发展湖底水质遥感反射率精确计算模型;4)发展适用于湖泊水体区域性Ⅱ类水体大气校正方法,并集成反演、遥感产品制作、分发等技术,构建湖泊水体水质/水色业务化运行体系.     

青藏高原湖泊沉积物与水体细菌群落共发生网络和群落构建过程差异解析

研究高原湖泊沉积物和水体细菌群落结构的差异对于水生态治理与保护具有重要的科学意义。本文选取青藏高原中东部共19个湖泊为研究对象,采集沉积物和水体样本,使用高通量测序技术和统计分析,解析湖泊沉积物和水体细菌群落结构、空间分布特征、共发生网络模式、群落构建过程的差异及其原因。研究结果表明沉积物细菌群落alpha多样性指数(Chao1指数、Shannon多样性指数、Faith’s系统发育指数和Pielou均匀度指数)均显著高于水体细菌群落。沉积物和水体细菌的优势门均为Proteobacteria(相对丰度为45.33%和41.77%),而水体细菌群落结构具有较强的空间异质性。主坐标分析和相似性分析结果表明沉积物和水体中细菌群落结构差异极显著,海拔和纬度分别是青藏高原湖泊沉积物和水体细菌群落结构变化最显著的驱动因子。沉积物和水体细菌群落均具有明显的距离衰减模式,而沉积物细菌比水体细菌群落在空间上的周转率小。采样湖泊沉积物细菌群落共发生网络复杂度和稳定性高于水体细菌。沉积物和水体细菌群落的平均归一化随机率值分别为0.24和0.72,表明确定性过程在沉积物细菌群落构建的过程中起到了重要的作用,而水...     

Empirical mass balance modelling of South American tropical glaciers: case study of Antisana volcano,Ecuador

Abstract

Most Latin American glaciers are located in the tropical Andes. The melting processes of Glacier “15” on Antisana volcano were studied to understand the relationship between glacier retreat and natural climate variability and global climate change. Glaciers on the Antisana volcano are crucial sources of water as they feed the headwater rivers that supply Quito with potable water. The aim of this study was to build empirical models based on multiple correlations to reconstruct the mass loss of glaciers over a period of 10 years at three scales: local (data recorded by meteorological stations located around the volcano), regional (data from stations located around the country) and global (re-analysis data). Data quality was checked using graphical and statistical methods. Several empirical models based on multiple correlations were created to generate longer time series (42 and 115 years) of the mass balance for the glacier ablation zone. The long mass balance series were compared with the temperature variation series of the Earth’s surface in the Southern Hemisphere to estimate the relation between the mass balance and global warming. Our results suggest that the meteorological factors that best correlate with mass balance are temperature and wind.

Editor D. Koutsoyiannis     

Describing alpine lake influence on stream network temperatures: A statistical modelling approach

Systematic variations in atmospheric heat exchange, surface residence time, and groundwater influx across montane stream networks commonly produce an increasing stream temperature trend with decreasing elevation. However, complex stream temperature profiles that differ from this common longitudinal trend also exist, suggesting that stream temperatures may be influenced by complex interactions among hydrologic and atmospheric processes. Lakes within stream networks form one potential source of temperature profile complexity due to the spatially variable contribution of lake-sourced water to stream flow. We investigated temperature profile complexity in a multi-season stream temperature dataset collected across a montane stream network containing many alpine lakes. This investigation was performed by making comparisons between multiple statistical models that used different combinations of stream and lake characteristics to represent specific hypotheses for the controls on stream temperature. The compared models included a set of models which used a topographically derived estimate of the hydrologic influence of lakes to separate and quantify the effects of stream elevation and lake source-water contributions to longitudinal stream temperature patterns. This source-water mixing model provided a parsimonious explanation for complex stream-network temperature patterns in the summer and autumn, and this approach may be further applicable to other systems where stream temperatures are influenced by multiple water sources. Simpler models that discounted lake effects were more optimal during the winter and spring, suggesting that complex patterns in stream temperature profiles may emerge and subside temporally, across seasons, in response to diversity of water temperatures from different sources.     

The maximum depth of a river as a stochastic process

The maximum depth of a river section is schematized as a non‐stationary continuous‐parameter continuous stochastic process, with a three‐parameter lognormal distribution. Two processes, represented by a first‐order and a second‐order differential equation, are considered. Non‐stationarity is accounted for by the mean, the other parameters being assumed constant. The continuous processes are then discretized as AR(1) and ARMA(2,1) processes respectively, and used for computing the conditional probability (which is of practical interest) for a given maximum depth not to be exceeded in a period of given length. The models are applied to the River Po (Italy) and the AR(1) model is found to be preferable. An analysis of the effect of discretizing the parameter is also carried out, considering the second‐order model and the conditional probability, for which analytical results for the continuous‐parameter model are available. Copyright © 2002 John Wiley & Sons, Ltd.     

Analysis of recharge-induced geochemical change in a contaminated aquifer

Recharge events that deliver electron acceptors such as O2, NO3, SO4, and Fe3+ to anaerobic, contaminated aquifers are likely important for natural attenuation processes. However, the specific influence of recharge on (bio)geochemical processes in ground water systems is not well understood. The impact of a moderate-sized recharge event on ground water chemistry was evaluated at a shallow, sandy aquifer contaminated with waste fuels and chlorinated solvents. Multivariate statistical analyses coupled with three-dimensional visualization were used to analyze ground water chemistry data (including redox indicators, major ions, and physical parameters) to reveal associations between chemical parameters and to infer processes within the ground water plume. Factor analysis indicated that dominant chemical associations and their interpreted processes (anaerobic and aerobic microbial processes, mineral precipitation/dissolution, and temperature effects) did not change significantly after the spring recharge event of 2000. However, the relative importance of each of these processes within the plume changed. After the recharge event, the overall importance of aerobic processes increased from the fourth to the second most important factor, representing the variability within the data set. The anaerobic signatures became more complex, suggesting that zones with multiple terminal electron-accepting processes (TEAPs) likely occur in the same water mass. Three-dimensional visualization of well clusters showed that water samples with similar chemical associations occurred in distinct water masses within the aquifer. Water mass distinctions were not based on dominant TEAPs, suggesting that the recharge effects on TEAPs occurred primarily at the interface between infiltrating recharge water and the aquifer.     

A conceptually based stochastic point process model for daily stream‐flow generation

Based on the stochastic and phenomenological aspects of hydrological processes, a conceptually based stochastic point process (SPP) model for daily stream‐flow generation is proposed in this paper. In which, storms are defined by a stochastic point process with marked values. All the random variables defining the process are assumed to be mutually independent, which constitutes a compound Poisson point process. The direct surface runoff is regarded as occurring from storage in a cascade of surface linear reservoirs and is responsible for the short‐term variation of the daily stream flows. The baseflow component is considered as coming from subsurface/groundwater storage and is responsible for the long‐term persistence of the storm time‐series. This type of model is proposed as a more realistic model of daily stream flow than models based on pure stochastic processes. Studies on the instantaneous unit hydrograph and the mechanism of baseflow could thereby provide some parameters for this model. Copyright © 2002 John Wiley & Sons, Ltd.     

Prediction of dissolved inorganic nitrogen (DIN) concentrations in deep,seasonally stratified lakes based on rates of DIN input and N removal processes

Mean dissolved inorganic nitrogen concentrations ([DIN]) in deep, seasonally stratified lakes with comparable DIN inputs can differ by up to a factor of 3 depending on hydraulic and morphometric properties and/or different trophic states of the lakes. In such lakes, net N sedimentation rates were estimated with two independent methods (sediment core analysis and input-output mass balances). They were higher in eutrophic lakes (Mean: 5.1; SD: ± 1.6 g m^–2 yr^–1; n = 13) than in oligotrophic lakes (1.6 ± 1.0 g m^–2 yr^–1; n = 3), but independent of [DIN]. Gaseous N loss rates to the atmosphere, as calculated from combined N- and P-mass balances from selected lakes, ranged from 0.9 to 37.4 g m^–2 yr^–1 (n = 10) and were positively correlated with [DIN]. Reduction of NO ₃ ^- to N₂ is assumed to be the main cause for gaseous N losses. A simple one-box mass balance model for [DIN], based on DIN input and rates and kinetics of N removal processes (net sedimentation and gaseous N loss) is proposed, and validated with a data base on [DIN] and DIN input in 19 deep, seasonally stratified lakes of central Europe. The model illustrated that the amount of water loading per unit surface area of a lake (called water discharge height q) is the critical parameter determining mean lake [DIN] relative to mean input [DIN]. Lakes with a q > 50 m yr^–1 have average [DIN] similar to the [DIN] of the inflows regardless of their trophic states, because input and outflow exceed lake-internal N removal processes. A high primary production favors DIN removal in lakes with q < 50 m yr^–1. It is concluded that measures to decrease primary production, e.g. by means of P removal programs, lead to an increase of [DIN] in lakes.     

Development of spatial regression models for predicting summer river temperatures from landscape characteristics: Implications for land and fisheries management

There is increasing demand for models that can accurately predict river temperature at the large spatial scales appropriate to river management. This paper combined summer water temperature data from a strategically designed, quality controlled network of 25 sites, with recently developed flexible spatial regression models, to understand and predict river temperature across a 3,000 km² river catchment. Minimum, mean and maximum temperatures were modelled as a function of nine potential landscape covariates that represented proxies for heat and water exchange processes. Generalised additive models were used to allow for flexible responses. Spatial structure in the river network data (local spatial variation) was accounted for by including river network smoothers. Minimum and mean temperatures decreased with increasing elevation, riparian woodland and channel gradient. Maximum temperatures increased with channel width. There was greater between‐river and between‐reach variability in all temperature metrics in lower‐order rivers indicating that increased monitoring effort should be focussed at these smaller scales. The combination of strategic network design and recently developed spatial statistical approaches employed in this study have not been used in previous studies of river temperature. The resulting catchment scale temperature models provide a valuable quantitative tool for understanding and predicting river temperature variability at the catchment scales relevant to land use planning and fisheries management and provide a template for future studies.