Search

Highlights • More than 150 papers dealing with applications of GP in water resource engineering were reviewed. • Fundamentals of GP variants were provided and compared with each other. • Selection ways of GP parameters were discussed. • Future scope of studies were demonstrated. Abstract The state-of-the-art genetic programming (GP) method is an evolutionary algorithm for automatic generation of computer programs. In recent decades, GP has been frequently applied on various kind of engineering problems and undergone speedy advancements. A number of studies have demonstrated the advantage of GP to solve many practical problems associated with water resources engineering (WRE). GP has a unique feature of introducing explicit models for nonlinear processes in the WRE, which can provide new insight into the understanding of the process. Considering continuous growth of GP and its importance to both water industry and academia, this paper presents a comprehensive review on the recent progress and applications of GP in the WRE fields. Our review commences with brief explanations on the fundamentals of classic GP and its advanced variants (including multigene GP, linear GP, gene expression programming, and grammar-based GP), which have been proven to be useful and frequently used in the WRE. The representative papers having wide range of applications are clustered in three domains of hydrological, hydraulic, and hydroclimatological studies, and outlined or discussed at each domain. Finally, this paper was concluded with discussions of the optimum selection of GP parameters and likely future research directions in the WRE are suggested. [ABSTRACT FROM AUTHOR]

Summary The multisensor Quantitative Precipitation Estimates (MQPEs) created by the US National Weather Service (NWS) are subject to a non-stationary bias. This paper quantifies the impacts of climatological adjustment of MQPEs alone, as well as the compound impacts of adjustment and model calibration, on the accuracy of simulated flood peak magnitude and that in detecting flood events. Our investigation is based on 19 watersheds in the mid-Atlantic region of US, which are grouped into small (<500 km 2 ) and large (>500 km 2 ) watersheds. NWS archival MQPEs over 1997–2013 for this region are adjusted to match concurrent gauge-based monthly precipitation accumulations. Then raw and adjusted MQPEs serve as inputs to the NWS distributed hydrologic model-threshold frequency framework (DHM-TF). Two experiments via DHM-TF are performed. The first one examines the impacts of adjustment alone through uncalibrated model simulations, whereas the second one focuses on the compound effects of adjustment and calibration on the detection of flood events. Uncalibrated model simulations show broad underestimation of flood peaks for small watersheds and overestimation those for large watersheds. Prior to calibration, adjustment alone tends to reduce the magnitude of simulated flood peaks for small and large basins alike, with 95% of all watersheds experienced decline over 2004–2013. A consequence is that a majority of small watersheds experience no improvement, or deterioration in bias (0% of basins experiencing improvement). By contrast, most (73%) of larger ones exhibit improved bias. Outcomes of the detection experiment show that the role of adjustment is not diminished by calibration for small watersheds, with only 25% of which exhibiting reduced bias after adjustment with calibrated parameters. Furthermore, it is shown that calibration is relatively effective in reducing false alarms (e.g., false alarm rate is down from 0.28 to 0.19 after calibration for small watersheds with calibrated parameters); but its impacts on detection rate are mixed. As an example, the detection rate of 2-Y events in fact declines for small watersheds after calibration is performed (from 0.4 to 0.28, and from 0.28 to 0.19 with raw and adjusted MQPE, respectively). These mixed outcomes underscore the complex interplays between errors in MQPEs, conditional bias in the reference gauge-based analysis, and structural deficiencies of the hydrologic model. [ABSTRACT FROM AUTHOR]

Summary One of the fundamental steps in regional rainfall frequency analysis is deciding the method by which rainfall stations are to be grouped together to form regions. This paper compares three methods of forming regions for use in estimating design rainfalls: a fixed region approach where all the available sites are included in a single region, a Region of Influence (ROI) approach based on geographical proximity and a hybrid approach where sites with similar topographic orientations are grouped together. The three region types were implemented in a Bayesian Generalized Least Squares Regression (BGLSR) framework which leads to regionalized regression equations that can be used to predict rainfall L-moments at ungauged sites. A leave-one-out cross validation approach was used to compare the relative accuracy, reliability and uncertainty of the derived rainfall statistics and resulting estimates of the rainfall quantiles. The study used data from two areas of Australia chosen for their highly varied topography and different climatic influences. It was found that all three methods provided good estimates of the L-moment statistics and the rainfall quantiles. The hybrid approach produced the smallest errors in the South-East Queensland region whilst for the Tasmanian region the fixed region approach was best. The results from this study show that although there is a slight benefit in using the proposed hybrid approach for BGLSR, these benefits were minor compared to maximizing the number of stations used to calibrate the BGLSR equations. This conclusion regarding the number of stations could be tested in future work by repeating the analyses in areas with sparser station density. Another test could be to simulate reduced station coverage in the current study areas by leaving stations out of the analyses. Finally it would be interesting to see if similar results are obtained by expanding the study area so that different climatological regimes are assessed. [ABSTRACT FROM AUTHOR]

Summary: Piezometric variability was investigated using monthly time series from 40 piezometers in different areas across the Upper Normandy region, over the 1968–2009 period. The piezometers were classified according to their hydrological behaviour. This paper aimed at characterising the dominant modes explaining piezometric variability and at investigating the causes of this variability. Significant components in the piezometric signals were identified: the annual cycle and an interannual mode (4–12years) with different variances for each observation borehole group. The influence of the morphostructural organisation of the Upper Normandy chalk plateau on the modalities of the water level variability in the chalk aquifer was defined. The relative contributions and standard deviations of each component found in the signal time series of the groundwater level was estimated. The highest values were recorded in the Northern study area, and clay-with-flints and aquifer thicknesses influence the transfer on an annual scale, while the loess thickness influences the transfer on a multi-year scale. [ABSTRACT FROM AUTHOR]

Summary: Water companies in the United Kingdom have considered climate change in their water resources plans for more than a decade through studies funded by UK Water Industry Research (UKWIR). This paper presents an initial assessment of the impact of the UK Climate Projections 2009 (UKCP09) on river flows at a national scale for the 2020s under the A1B scenario and the implications for water resource planning. A daily hydrological modelling framework based on two conceptual model structures and the Generalized Likelihood Uncertainty Estimation (GLUE) methodology has been applied to 70 catchments across the UK. A Latin Hypercube Sampling approach was adopted to handle the probabilistic nature of UKCP09. Results show a decrease in mean annual flow over most of the UK, with negative median values of all monthly changes except in winter over the western and northern mountainous areas. Furthermore the results indicate a high likelihood of a significant decline in summer flows. An analysis of variance shows that the main uncertainty in river flow changes comes from the spread in climate projections. Finally results are found to be consistent with a previous UKWIR assessment based on individual projections from 6 Global Climate Models (GCMs) under the A2 scenario. The reduction in summer low-flows, critical for water resources, appears however more limited with UKCP09. Although most expected changes are within natural variability, the drier conditions overall and the greater spread of results with respect to previous assessments indicates a need for testing the robustness of water resource plans. [Copyright &y& Elsevier]

Summary: This paper presents an analysis on the hydrological regimes from several rivers flowing from the western slope of the Andes Cordillera. We compiled a database of 40 unimpaired average monthly streamflow records located in central and southern Chile, spanning the region between latitudes 30°S and 40°S and the period between years 1961 and 2006. River categories encompass snowmelt, mixed and rainfall dominated watersheds. The water year hydrograph center of timing (CT) is used as an indicator of flow regime and as a proxy for streamflow distribution throughout the year. The CT for each river is correlated to local climatic variables such as seasonal temperature and precipitation, and to large-scale circulation indexes. We assess CT trends using the Mann–Kendall test for varying time windows within the study period. Overall results for the 1961–2006 period show a significant (95% confidence level) negative trend (CT date shifting towards earlier in the year) for 23 out of the 40 analyzed series. However, when studying time windows within the entire period there seems to be a break in the trend during the late 1970s, with positive or null changes for the latter half of the study period for most of the stations analyzed. We found that even though the CT is significantly correlated (r >0.5) to temperature indexes in the snowfed watersheds, there has been no significant change in CT timing for most of these watersheds despite documented warming on the region. CT for all types of watersheds present clear correlation to precipitation indexes (r >0.8), and we observed high and significant correlation between CT and ONI, SOI and MEI indexes for all the study region: higher (lower) precipitation amounts during El Niño (La Niña), rather than temperature variations, seem to be the most important factor controlling CT timing during these episodes. In order of importance, the CT showed higher correlation with annual precipitation amounts and timing, MEI, ONI and SOI, mean winter and spring temperature, and finally PDO. The detected trends are consistent only with trends in precipitation, as rivers that showed high correlation to temperature do not show any trend. These conclusions suggest that rivers on the region show a different response to temperature changes when compared to rivers of similar regime in the northern hemisphere, and that precipitation amount and timing are the source of most of the variability in streamflow timing for both snowmelt and rainfall dominated watersheds. Temperature induced variability in streamflow timing should be studied with more precise hydrological models, as the CT has a complex behavior due to multiple climatical dependencies. Studying such behavior using only historical data might be insufficient for drawing more precise conclusions. [Copyright &y& Elsevier]

Summary: This paper presents that a simple regional regression-based approach with readily available geographic and meteorological parameters as predictors could be a viable method for mapping snow water equivalent (SWE) climatology in the mountainous areas of the western United States. Such climatological information is potentially useful for several hydrologic applications, including estimation of real-time SWE grids and calibration and evaluation of SWE estimates obtained from remote sensing or through various snow models. Regional delineation for the mountainous regions of the western United States was done through cluster analysis in consideration of the characteristics of seasonal snow pack accumulation and ablation processes. Various geographic and meteorological parameters were further investigated through stepwise regression as potential predictors of monthly changes in climatological SWE in each delineated region. In-situ measurements of SWE, obtained from the Natural Resources Conservation Service’s Snow Telemetry network for the period from 1980 to 2004, were used to calibrate regional equations. The spatial and temporal resolution of the analysis was based on the resolution of available meteorological data – 4km and 1month, respectively. For a monthly time step, the reliability of the SWE estimates did not significantly increase when the number of regions was more than five. The performance of the developed regional equations was evaluated via a jack-knife technique. The regional equations developed using monthly resolution and large regions provided reliable estimates for the majority of regions from October to March, but not in April, especially in the North Pacific and Southwest regions. Sub-regionalization, a finer time step for the analysis and/or inclusion of additional SWE predictor variables may increase the reliability of SWE climatological estimates for warmer months. [ABSTRACT FROM AUTHOR]

Summary: The purpose of the paper is to show that the groundwater impacts of climatic cycles vary spatially across the aquifer depending on local hydrogeologic characteristics. Spectral analysis was applied 53 piezometers to study cycles patterns of variability in time and space in the Vega de Granada aquifer. Other variables like rainfall data, flows of rivers, and temperature were also analysed in order to determine the influence of climatological cycles. A number of long-term cycles were detected in the data series, with a time evolution that was well reflected in most piezometers. Many of the piezometers have been registering the position of the hydraulic head on a monthly basis for the past 30 years. The climatological time series represent nearly 50 years of monthly recordings, while the time series of river flow covers more than 30 years. Using spectral methods, we distinguished four different cycles in the time evolution of the aquifer hydraulic head: a decadal cycle (peaks between 8 and 11 years), a 3.2 year cycle, an annual one, and a semi-annual period. We also detected a 22 year cycle in a rainfall time series, though with less clarity. Whereas the annual cycle is ubiquitous, reflecting the annual hydrological cycle, the decadal and 3.2 year periods show a much more interesting spatial distribution. The decadal cycle can certainly be traced to hydrological variations involving a climatological regimen, probably induced by North Atlantic Oscillation (NAO). Groundwater impacts of climatic cycles vary spatially across the aquifer depending on local hydrogeologic characteristics. Specifically, the piezometers located close to streamflows reaching the aquifer, or along the main river in those areas where it recharges the aquifer, present a statistically significant decadal cycle. NAO wet periods imply more rainfall than average over Spain and thus more recharge into the aquifer increasing the water table. The opposite occurs during the dry periods. The presence and spatial distribution of the 3-year cycle would have an explanation similar to that of the decadal cycle, though it is identified at fewer points of piezometric control. [Copyright &y& Elsevier]

Summary: The knowledge of soil moisture spatial variability is an important issue for hydrological and climatic studies. This paper investigates the soil moisture spatial pattern obtained from spot measurements carried out in different experimental plots located in small basins of Central Italy. The near-surface volumetric soil moisture content was measured using a portable time domain reflectometer (TDR) over regular grids whose areas ranging in extension from 9 to 8800m2. The sampling in each site was periodically repeated in order to investigate the moisture pattern as function of wetness conditions. The statistical properties of the measurements and the relation of the statistical parameters with the field-mean soil moisture content were analyzed. The results states that the probability distribution can be assumed normal for soil moisture in flat areas and for normalized soil moisture in gentle slope areas. A general decreasing trend of variance with increasing mean moisture content was observed. Moreover, considering the relationship between the field-mean water content and the coefficient of variation an optimal measurements number ranging from 15 to 35, in the worse case, was found for the study areas. The higher number was required for the site with more significant relief. These results were also confirmed through the geostatistical analysis and the correlation with the topographical attributes. Specifically, the correlation lengths were found increasing with the mean areal slope and the mean drainage direction and ranging between 10 and 15m. This is corroborated through geostatistical analysis showing that the natural logarithm of the specific upslope area was the best spatial predictor for near-surface soil moisture when the water volume in the soil is higher. On this basis, the surface and subsurface lateral redistribution of water can be surmised fundamental in spatial organization of soil moisture. [Copyright &y& Elsevier]

Abstract: The perception of a changing climate, which impacts also hydrological processes, is now generally admitted. However, the way of handling the changing nature of climate in hydrologic practice and especially in hydrological statistics has not become clear so far. The most common modelling approach is to assume that long-term trends, which have been found to be omnipresent in long hydrological time series, are ‘deterministic’ components of the time series and the processes represented by the time series are nonstationary. In this paper, it is maintained that this approach is contradictory in its rationale and even in the terminology it uses. As a result, it may imply misleading perception of phenomena and estimate of uncertainty. Besides, it is maintained that a stochastic approach hypothesizing stationarity and simultaneously admitting a scaling behaviour reproduces climatic trends (considering them as large-scale fluctuations) in a manner that is logically consistent, easy to apply and free of paradoxical results about uncertainty. [Copyright &y& Elsevier]

Abstract: This paper deals with the effects of land levelling on soil hydrological properties and spatial variability of soil moisture at different depths in bare vineyard fields under different rainfall distributions over the year. The research was carried out in a levelled vineyard in the Penedès region (NE Spain), which is representative of the land management practices in that area. Twelve locations were instrumented in the selected plot to control soil moisture periodically at different depths (0–20, 20–40, 40–60cm, using a TDR tube probe), and runoff (using Gerlach collectors located at the same positions) during the period April 2002–March 2004. Rainfalls were recorded at one-minute intervals in the experimental field using a pluviometer linked to a data-logger. The vine yield and its relationship with the rainfall distribution over the year was evaluated in this levelled vineyard vs. a less transformed vineyard. The land levelling works carried out in the plot before the establishment of the vineyard led to major differences in soil depth (ranging from 50 to 110cm) and in soil layer characteristics along and across the plot. This helps to generate different soil moisture conditions at the same depth at different positions of the plot. The areas located at the top of the plot showed soil moisture conditions in the surface layer that were up to 5% lower than down slope, although with high variability across the plot. Soil moisture in the whole profile depends on the intensity of rainfall more than total rainfall. At high intensities, soil moisture increases only occur in the surface layer, but no significant increases are observed in deeper layers, where vines have most of their roots. On the other hand, under low intensity rainfalls the soil water content increases in the whole profile. Levelled plots showed a yield reduction in comparison with those undergoing less transformation, which was exacerbated by a very irregular rainfall distribution: for a similar total rainfall, but with dry periods during the stages in which the water requirement for vines is higher, grape production was up to 50% lower in areas with high land transformation, where Chardonnay was planted, and up to 16% lower at the bottom areas, where Macabeo was planted. [Copyright &y& Elsevier]