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.     

Hydrologic processes at the urban residential scale

In the face of increasing urbanization, there is growing interest in application of microscale hydrologic solutions to minimize storm runoff and conserve water at the source. In this study, a physically based numerical model was developed to understand hydrologic processes better at the urban residential scale and the interaction of these processes among different best management practices (BMPs). This model simulates hydrologic processes using an hourly interval for over a full year or for specific storm events. The model was applied to treatment and control single‐family residential parcels in Los Angeles, California. Data collected from the control and treatment sites over 2 years were used to calibrate and validate the model. Annual storm runoff to the street was eliminated by 97% with installation of rain gutters, a driveway interceptor, and lawn retention basin. Evaluated individually, the driveway interceptor was the most effective BMP for storm runoff reduction (65%), followed by the rain gutter installation (28%), and lawn converted to retention basin (12%). An 11 m³ cistern did not substantially reduce runoff, but provided 9% of annual landscape irrigation demand. Simulated landscape irrigation water use was reduced 53% by increasing irrigation system efficiency, and adjusting application rates monthly based on plant water demand. The model showed that infiltration and surface runoff processes were particularly sensitive to the soil's physical properties and its effective depth. Replacing the existing loam soil with clay soil increased annual runoff discharge to the street by 63% when climate and landscape features remained unchanged. Copyright © 2006 John Wiley & Sons, Ltd.     

Modelling the potential impacts of groundwater hydrology on long‐term drainage basin evolution

Fluvial erosion processes are driven by water discharge on the land surface, which is produced by surface runoff and groundwater discharge. Although groundwater is often neglected in long‐term landscape evolution problems, water table levels control patterns of Dunne runoff production, and groundwater discharge can contribute significantly to storm flows. In this analysis, we investigate the role that groundwater movement plays in long‐term drainage basin evolution by modifying a widely used landscape evolution model to include a more detailed representation of basin hydrology. Precipitation is generated by a stochastic process, and the precipitation is partitioned between surface runoff and groundwater recharge using a specified infiltration capacity. Groundwater flow is simulated by a dynamic two‐dimensional Dupuit equation for an unconfined aquifer with an irregular underlying impervious layer. The model is applied to the WE‐38 basin, an experimental catchment in Pennsylvania, because 60–80 per cent of the discharge is derived from groundwater and substantial hydrologic and geomorphic information is available. The hydrologic model is first calibrated to match the observed streamflows, and then the combined hydrologic/geomorphic model is used to simulate scenarios with different infiltration capacities. The results of this modelling exercise indicate that the basin can be divided into three zones with distinct streamflow‐generating characteristics, and different parts of the basin can have different geomorphic effective events. Over long periods of time, scenarios in which groundwater discharge is large tend to modify the topography in a way that promotes groundwater discharge and inhibits Dunne runoff. Copyright © 2006 John Wiley & Sons, Ltd.     

Evaluation of distributed BMPs in an urban watershed—High resolution modeling for stormwater management

The urban environment modifies the hydrologic cycle resulting in increased runoff rates, volumes, and peak flows. Green infrastructure, which uses best management practices (BMPs), is a natural system approach used to mitigate the impacts of urbanization onto stormwater runoff. Patterns of stormwater runoff from urban environments are complex, and it is unclear how efficiently green infrastructure will improve the urban water cycle. These challenges arise from issues of scale, the merits of BMPs depend on changes to small‐scale hydrologic processes aggregated up from the neighborhood to the urban watershed. Here, we use a hyper‐resolution (1 m), physically based hydrologic model of the urban hydrologic cycle with explicit inclusion of the built environment. This model represents the changes to hydrology at the BMP scale (~1 m) and represents each individual BMP explicitly to represent response over the urban watershed. Our study varies both the percentage of BMP emplacement and their spatial location for storm events of increasing intensity in an urban watershed. We develop a metric of effectiveness that indicates a nonlinear relationship that is seen between percent BMP emplacement and storm intensity. Results indicate that BMP effectiveness varies with spatial location and that type and emplacement within the urban watershed may be more important than overall percent.     

Seasonal dynamics in shallow freshwater pond‐peatland hydrochemical interactions in a subarctic permafrost environment

Terrestrial and aquatic ecological productivity are often nutrient limited in subarctic permafrost environments. High latitude regions are experiencing significant climatic change, including rapid warming and changing precipitation patterns, which may result in changes in nutrient dynamics within terrestrial and aquatic systems and hydrochemical transport between them. The objective of this research was to characterize changes in runoff quantity and quality within, and between peatlands and ponds throughout the snow‐free summer season. Two ponds and their catchments were monitored over the snow‐free season to measure changes in hydrologic storage, and to determine how water chemistry changed with the evolution of the frost table depth. Thresholds in hydrologic storage combined with frost table position (which inhibited infiltration and storage) produced nonlinear responses for runoff generation through highly conductive shallow peat layers while deeper, less conductive layers retarded flow. Greater inputs were required to exceed hydrologic storage (fill and spill) as a deepening frost table increased the hydrologically active portion of the soil, leading to seasonal variability in runoff pathways between peatlands and ponds. Runoff contributions to ponds were an integral component of the snow‐free water balance during the study period, contributing up to 60% of all snow‐free inputs. Groundwater chemistry (and pond chemistry following runoff events when ponds were connected with peatlands) reflected the different depths of peat and mineral soil accessed throughout the season. This work has improved scientific understanding of the combined controls of hydrologic inputs and ground frost on runoff and nutrient transport between peatlands and ponds, and sheds insight into how nutrient dynamics in cold regions may evolve under a changing climate.     

Difference infiltrometer: a method to measure temporally variable infiltration rates during rainstorms

We developed a difference infiltrometer to measure time series of non‐steady infiltration rates during rainstorms at the point scale. The infiltrometer uses two, tipping bucket rain gages. One gage measures rainfall onto, and the other measures runoff from, a small circular plot about 0.5‐m in diameter. The small size allows the infiltration rate to be computed as the difference of the cumulative rainfall and cumulative runoff without having to route water through a large plot. Difference infiltrometers were deployed in an area burned by the 2010 Fourmile Canyon Fire near Boulder, Colorado, USA, and data were collected during the summer of 2011. The difference infiltrometer demonstrated the capability to capture different magnitudes of infiltration rates and temporal variability associated with convective (high intensity, short duration) and cyclonic (low intensity, long duration) rainstorms. Data from the difference infiltrometer were used to estimate saturated hydraulic conductivity of soil affected by the heat from a wildfire. The difference infiltrometer is portable and can be deployed in rugged, steep terrain and does not require the transport of water, as many rainfall simulators require, because it uses natural rainfall. It can be used to assess infiltration models, determine runoff coefficients, identify rainfall depth or rainfall intensity thresholds to initiate runoff, estimate parameters for infiltration models, and compare remediation treatments on disturbed landscapes. The difference infiltrometer can be linked with other types of soil monitoring equipment in long‐term studies for detecting temporal and spatial variability at multiple time scales and in nested designs where it can be linked to hillslope and basin‐scale runoff responses. Published 2012. This article is a U.S. Government work and is in the public domain in the USA.     

Impact of an unsealed forest road stream crossing: water quality and sediment sources

Turbidity monitoring and rainfall and runoff simulation experiments were conducted at a newly constructed unsealed road stream crossing to determine the quantity and sources of sediment entering the stream. Continuous measurements of turbidity and estimation of total suspended solids (TSS) concentration upstream and downstream of the stream culvert were taken over a 5 month period. There was a statistically significant difference in turbidity and TSS downstream of the crossing during baseflow conditions, but the quality of the water column remained good during non‐rain periods. Rainfall events comprised around 20% of the observation period and led to decreases in water quality downstream of the crossing. Water quality could be considered as degraded for 10% of the observations. This was during a period when the rainfall was 65% of the long‐term average. Calculated suspended sediment loads were 0·78 t upstream and 2·77 t downstream, an increase of 3·5. It was estimated that at least 2–3 t of bedload material was also added to the stream during the crossing construction and from subsequent erosion. This material is a deposit on the cobble stream bed, and is most likely to degrade aquatic ecosystem values. Rainfall and runoff simulation revealed the principal sediment sources to be a fillslope that contributed coarse bedload material through rill erosion and unprotected toe scour, and the unmetalled road verge that provided fines. Although the quality of water column was good for the majority of the observations, the new Australian and New Zealand Water Quality Guidelines for Fresh and Marine Waters suggest this site exceeded ‘trigger levels’ that would warrant further investigation for both the water column and the bed deposits. Copyright © 2002 John Wiley & Sons, Ltd.     

Identification of physical processes inherent in artificial neural network rainfall runoff models

The emergence of artificial neural network (ANN) technology has provided many promising results in the field of hydrology and water resources simulation. However, one of the major criticisms of ANN hydrologic models is that they do not consider/explain the underlying physical processes in a watershed, resulting in them being labelled as black‐box models. This paper discusses a research study conducted in order to examine whether or not the physical processes in a watershed are inherent in a trained ANN rainfall‐runoff model. The investigation is based on analysing definite statistical measures of strength of relationship between the disintegrated hidden neuron responses of an ANN model and its input variables, as well as various deterministic components of a conceptual rainfall‐runoff model. The approach is illustrated by presenting a case study for the Kentucky River watershed. The results suggest that the distributed structure of the ANN is able to capture certain physical behaviour of the rainfall‐runoff process. The results demonstrate that the hidden neurons in the ANN rainfall‐runoff model approximate various components of the hydrologic system, such as infiltration, base flow, and delayed and quick surface flow, etc., and represent the rising limb and different portions of the falling limb of a flow hydrograph. Copyright © 2004 John Wiley & Sons, Ltd.     

Modelling of stormwater biofilters under random hydrologic variability: a case study of a car park at Monash University,Victoria (Australia)

Biofiltration systems represent an effective technology for the management of urban stormwater runoff volumes and quality. The performance of these systems, although largely dependent on their physical characteristics, is also strongly affected by the natural variability of runoff occurrence and volumes. This article presents a model that describes the statistical behaviour of the main variables involved in the water balance of a biofiltration system, given its main physical properties (filter media and vegetation types) and accounting for the natural inflow variability in terms of occurrence and water volumes. The model permits the analytical derivation of the long‐term (e.g. annual) probability density function of the soil water content in the filter media and the estimation of the main statistics of water fluxes, that is, outflow, evapotranspiration and overflow. By relating the soil water content in the filter media before inflow events to the outflow total nitrogen concentrations, the model also gives estimates of the statistics of nitrogen removal performance as a function of inflow variability. The model was tested against field data collected at a stormwater biofiltration system in Melbourne, Australia. The model could be used to rapidly assess the hydrologic and nitrogen treatment performance of alternative applications of biofiltration for stormwater management across a range of climates. Copyright © 2011 John Wiley & Sons, Ltd.     

Annual runoff assessment in arid and semiarid Mediterranean watersheds under the Budyko's framework

The solution of many practical water problems is strictly connected to the availability of reliable and widespread information about runoff. The estimation of mean annual runoff and its interannual variability for any basin over a wide region, even if ungauged, would be fundamental for both water resources assessment and planning and for water quality analysis. Starting from these premises, the main aim of this work is to show a new approach, based on the Budyko's framework, for mapping the mean annual surface runoff and deriving the probability distribution of the annual runoff in arid and semiarid watersheds. As a case study, the entire island of Sicily, Italy, is here proposed. First, time series data of annual rainfall, runoff, and reconstructed series of potential evapotranspiration have been combined within the Budyko's curve framework to obtain regional rules for rainfall partitioning between evapotranspiration and runoff. Then this knowledge has been used to infer long‐term annual runoff at the point scale by means of interpolated rainfall and potential evapotranspiration. The long‐term annual runoff raster layer has been obtained at each pixel of the drainage network, averaging the upstream runoff using advanced spatial analysis techniques within a GIS environment. Furthermore, 2 alternative methods are here proposed to derive the distribution of annual runoff, under the assumption of negligible interannual variations of basin water storage. The first method uses Monte Carlo simulations, combining rainfall and potential evapotranspiration randomly extracted from independent distributions. The second method is based on a simplification of the Budyko's curve and analytically provides the annual runoff distribution as the derived distribution of annual rainfall and potential evapotranspiration. Results are very encouraging: long‐term annual runoff and its distribution have been derived and compared with historical records at several gauged stations, obtaining satisfactory matching.     

Delineating runoff processes and critical runoff source areas in a pasture hillslope of the Ozark Highlands

The identification of runoff contributing areas would provide the ideal focal points for water quality monitoring and Best Management Practice (BMP) implementation. The objective of this study was to use a field‐scale approach to delineate critical runoff source areas and to determine the runoff mechanisms in a pasture hillslope of the Ozark Highlands in the USA. Three adjacent hillslope plots located at the Savoy Experimental Watershed, north‐west Arkansas, were bermed to isolate runoff. Each plot was equipped with paired subsurface saturation and surface runoff sensors, shallow groundwater wells, H‐flumes and rain gauges to quantify runoff mechanisms and rainfall characteristics at continuous 5‐minute intervals. The spatial extent of runoff source areas was determined by incorporating sensor data into a geographic information‐based system and performing geostatistical computations (inverse distance weighting method). Results indicate that both infiltration excess runoff and saturation excess runoff mechanisms occur to varying extents (0–58% for infiltration excess and 0–26% for saturation excess) across the plots. Rainfall events that occurred 1–5 January 2005 are used to illustrate the spatial and temporal dynamics of the critical runoff source areas. The methodology presented can serve as a framework upon which critical runoff source areas can be identified and managed for water quality protection in other watersheds. Copyright © 2008 John Wiley & Sons, Ltd.     

On the effects of triangulated terrain resolution on distributed hydrologic model response

Distributed hydrologic models based on triangulated irregular networks (TIN) provide a means for computational efficiency in small to large‐scale watershed modelling through an adaptive, multiple resolution representation of complex basin topography. Despite previous research with TIN‐based hydrology models, the effect of triangulated terrain resolution on basin hydrologic response has received surprisingly little attention. Evaluating the impact of adaptive gridding on hydrologic response is important for determining the level of detail required in a terrain model. In this study, we address the spatial sensitivity of the TIN‐based Real‐time Integrated Basin Simulator (tRIBS) in order to assess the variability in the basin‐averaged and distributed hydrologic response (water balance, runoff mechanisms, surface saturation, groundwater dynamics) with respect to changes in topographic resolution. Prior to hydrologic simulations, we describe the generation of TIN models that effectively capture topographic and hydrographic variability from grid digital elevation models. In addition, we discuss the sampling methods and performance metrics utilized in the spatial aggregation of triangulated terrain models. For a 64 km² catchment in northeastern Oklahoma, we conduct a multiple resolution validation experiment by utilizing the tRIBS model over a wide range of spatial aggregation levels. Hydrologic performance is assessed as a function of the terrain resolution, with the variability in basin response attributed to variations in the coupled surface–subsurface dynamics. In particular, resolving the near‐stream, variable source area is found to be a key determinant of model behaviour as it controls the dynamic saturation pattern and its effect on rainfall partitioning. A relationship between the hydrologic sensitivity to resolution and the spatial aggregation of terrain attributes is presented as an effective means for selecting the model resolution. Finally, the study highlights the important effects of terrain resolution on distributed hydrologic model response and provides insight into the multiple resolution calibration and validation of TIN‐based hydrology models. Copyright © 2005 John Wiley & Sons, Ltd.     

Performance of grass and eucalyptus riparian buffers in a pasture catchment,Western Australia,part 1: riparian hydrology

Rainfall takes many flowpaths to reach a stream, and the success of riparian buffers in water quality management is significantly influenced by riparian hydrology. This paper presents results from hydrometric monitoring of riparian buffer hydrology in a pasture catchment. Runoff processes and riparian flowpaths were investigated on two planar hillslopes with regenerating grass and E. globulus buffers. Surface runoff and subsurface flows (A‐ and B‐horizons) were measured for 3 years using surface runoff collectors, subsurface troughs and piezometers. Water volumes moving through the riparian buffers via the measured flowpaths were ranked B‐horizon ? surface runoff ≈ A‐horizon. Runoff volumes through the B‐horizon troughs were an order of magnitude greater than those recorded for the most productive surface runoff plots or the A‐horizon troughs. Subsurface runoff and saturation‐excess overland flow (SOF) were limited to the winter months, whereas infiltration‐excess overland flow (IEOF) can occur all year round during intense storms. Surface runoff was recorded on 33 occasions, mostly during winter (late May–early October), and total annual surface runoff volumes collected by the 20 unconfined (2 m wide) runoff plots varied between > 80 and < 20 m³. Subsurface flow only occurred in winter, and the 6 m wide B‐horizon subsurface troughs flowed above 1 l s^?1 continuously, whereas the A‐horizon troughs flowed infrequently (<6 days per year). In summer, surface runoff occurred as IEOF during intense storms in the E. globulus buffer, but not in the grass buffer. Observations suggest that surface crusting reduced the soil's infiltration capacity in the E. globulus buffer. During winter, SOF and seepage were observed in both buffers, but subsurface flow through the B‐horizon was the dominant flowpath. Key hydrologic differences between the grass and tree buffers are the generation of IEOF in the E. globulus buffer during intense summer storms, and the smaller subsurface runoff volumes and fewer flow days in the E. globulus buffer. Low surface runoff volumes are likely to limit the potential of these buffers to filter pollutants from surface runoff. High subsurface flow volumes and saturated conductivities are also likely to limit the residence time of water in the subsurface domain. Based on their hydrologic performance, the key roles of riparian buffers in this landscape are likely to be displacing sediment and nutrient‐generating activities away from streams and stabilizing channel morphology. Copyright © 2006 John Wiley & Sons, Ltd.     

Alteration of hydrologic indicators for Korean catchments under CMIP5 climate projections

The change of hydrological regimes may cause impacts on human and natural system. Therefore, investigation of hydrologic alteration induced by climate change is essential for preparing timely proper adaptation to the changes. This study employed 24 climate projections from the Coupled Model Intercomparison Project Phase 5 (CMIP5) under Representative Concentration Pathway (RCP) 4.5 scenario. The climate projections were downscaled at a station‐spacing for seven Korean catchments by a statistical downscaling method that preserves a long‐term trend in climate projections. Using an ensemble of future hydrologic projections simulated by three conceptual rainfall‐runoff models (GR4J, IHACRES, and Sacramento models), we calculated Hydrologic Alteration Factors (HAFs) to investigate degrees of variations in Indicators of Hydrologic Alteration (IHAs) derived from the hydrologic projections. The results showed that the seven catchments had similar trend in terms of the HAFs for the 24 IHAs. Given that more frequent severe floods and droughts were projected over Korean catchments, sound water supply strategies are definitely required to adapt to the alteration of streamflow. A wide range of HAFs between rainfall‐runoff models for each catchment was detected by large variations in the magnitude of HAFs with the hydrologic models and the difference could be the hydrologic prediction uncertainty. There were no‐consistent tendency in the order of HAFs between the hydrologic models. In addition, we found that the alterations of hydrologic regimes by climate change are smaller as the size of catchment is larger. Copyright © 2016 John Wiley & Sons, Ltd.     

Land cover effects on runoff patterns in eastern Piedmont (USA) watersheds

Physiography and land cover determine the hydrologic response of watersheds to climatic events. However, vast differences in climate regimes and variation of landscape attributes among watersheds (including size) have prevented the establishment of general relationships between land cover and runoff patterns across broad scales. This paper addresses these difficulties by using power spectral analysis to characterize area‐normalized runoff patterns and then compare these patterns with landscape features among watersheds within the same physiographic region. We assembled long‐term precipitation and runoff data for 87 watersheds (first to seventh order) within the eastern Piedmont (USA) that contained a wide variety of land cover types, collected environmental data for each watershed, and compared the datasets using a variety of statistical measures. The effect of land cover on runoff patterns was confirmed. Urban‐dominated watersheds were flashier and had less hydrologic memory compared with forest‐dominated watersheds, whereas watersheds with high wetland coverage had greater hydrologic memory. We also detected a 10–15% urban threshold above which urban coverage became the dominant control on runoff patterns. When spectral properties of runoff were compared across stream orders, a threshold after the third order was detected at which watershed processes became dominant over precipitation regime in determining runoff patterns. Finally, we present a matrix that characterizes the hydrologic signatures of rivers based on precipitation versus landscape effects and low‐frequency versus high‐frequency events. The concepts and methods presented can be generally applied to all river systems to characterize multiscale patterns of watershed runoff. Copyright © 2013 John Wiley & Sons, Ltd.     

Long‐term runoff forecasting by combining hydrological models and meteorological records

Reservoir operation is generally based on the inflows of the upstream catchment of the reservoir. If the arriving inflows can be forecasted, that can benefit reservoir operation and management. This study attempts to construct a long‐term inflow‐forecasting model by combining a continuous rainfall–runoff model with the long‐term weather outlook from the Central Weather Bureau of Taiwan. The analytical results demonstrate that the continuous rainfall–runoff model has good inflow simulation performance by using 10‐day meteorological and inflow records over a 33‐year period for model calibration and verification. The long‐term inflow forecasting during the dry season was further conducted by combining the continuous rainfall–runoff model and the long‐term weather outlook, which was found to have good performance. Copyright © 2005 John Wiley & Sons, Ltd.     

Evaluating hydrologic performance of bioretention cells in shallow groundwater

Bioretention cells, which are generally effective in controlling surface runoff and recharging groundwater, have been widely adopted as low impact development practices. However, shallow groundwater has limited their implementation in some locations due to the potential problems of a reduction in surface runoff control, groundwater pollution, and continuous groundwater drainage through the underdrain. Many guidelines have established minimum requirements for the groundwater depth below bioretention cells, but they may not be optimized for certain environmental conditions and bioretention cell designs. This study made use of a variably saturated flow model to examine the hydrologic performance of a single bioretention cell in shallow groundwater with event‐based simulations, considering a wide range of initial groundwater depths, media and in situ soil types, surface runoff loads, and underdrain sizes. Performance indicators (e.g., runoff reduction, time for infiltrated water to reach the bioretention cell bottom and the groundwater table, and height and dissipation time of groundwater mound) were evaluated to examine the processes of runoff generation, the formation and dissipation of groundwater mounds, and the bioretention cell's performance in a shallow groundwater environment. The most influential factors were the initial groundwater depth, the hydraulic conductivity of the media soil, and the rainfall runoff load. With a deeper initial groundwater table, infiltrated water took longer to reach the bioretention cell bottom and groundwater table. Groundwater mounds, however, took longer to dissipate even though they were smaller. The groundwater quality can be better protected if relatively less‐permeable soil types (e.g., sandy loam) are used as the media, although it may compromise the performance in runoff quantity control. However, only very high surface runoff loads would cause concerns regarding a reduction in runoff quantity control and possible groundwater contamination due to the shallow groundwater. A distance of 1.5–3 m between the bioretention cell bottom and the groundwater table is generally sufficient. The results of this study could help to guide the planning and design of bioretention cells in areas of shallow groundwater.     

Design measures to mitigate the impact of shallow groundwater on hydrologic performance of permeable pavements

Permeable pavements (PPs) are widely implemented in urban areas to mimic natural hydrologic processes through enhancing infiltration, and reducing, delaying, and retaining surface runoff. However, its performance can be affected by shallow groundwater since high soil moisture may inhibit its infiltration and exfiltration. This study built a numerical model, which was calibrated and validated based on laboratory experiment data, to evaluate the water balance and retention of PP in shallow groundwater conditions. It assessed the impacts of shallow groundwater and the hydrologic effectiveness of different PP design measures (i.e., building a PP with a smaller storage depth, implementing an underdrain at different elevations, and installing an impermeable liner) on relieving the impacts. Shallower groundwater led to larger amounts of surface runoff and underdrain flow, and a higher chance of saturating the PP reservoir. The three design measures had both benefits and drawbacks in mimicking natural hydrologic cycle and retaining the performance of PP under extreme conditions (e.g., areas of very shallow groundwater tables and/or extreme rainfalls). A PP with a smaller storage depth resulted in less underdrain flow but was prone to saturation. It is, thus, more recommended for PP with more-permeable subsoils, which can avoid frequent pavement saturation. Although a shallower PP corresponds to a smaller storage volume and shorter hydraulic retention time, it can increase the applicability of PP to shallow groundwater areas, which is beneficial to the regional hydrologic environment. Installing an underdrain generated underdrain flow, which is a burden to the downstream drainage system. However, it significantly reduced the surface runoff and the chance of saturating the PP reservoir, which, thus, is more recommended for PP with less-permeable subsoils. Comparatively, elevating the underdrain is recommended in areas of shallow groundwater because it can reduce the frequency and amount of groundwater-induced underdrain flow. In addition, a higher underdrain together with an impermeable liner can create a storage depth, increase the retention duration, enhance exfiltration and evaporation without increasing the saturation risk.