Subsurface Return Flow Research Articles

Quantifying Stream Return Flow of Agricultural Water Using SWAT-MODFLOW-PADDY Model in Korean Paddy Fields

In many countries, the irrigation return flow focuses only on surface and subsurface flows. In contrast, South Korea adopts a broader approach, defining the stream return flow as encompassing both quick and delayed return flows, which include subsurface flow and deep percolation. This study proposes redefining the stream return flow to include only the subsurface return flow, excluding deep percolation. We quantified the subsurface return flow and deep percolation using the SWAT-MODFLOW-PADDY (SMP) model, confirming that the current definition overestimated the stream return flow in Korea. The results show that the subsurface return flow accounted for 20% to 60% of the total infiltration, with the remaining 40% to 80% contributing to deep percolation and groundwater recharge. These findings reveal significant regional variations in the subsurface return flow rates, underscoring the limitations of applying a uniform stream return flow rate. We propose that allocated management water, subsurface, and quick return flows should be the primary components considered in stream return flow calculations, as the current practice of including delayed return flow leads to overestimated results. This study highlights the challenges in monitoring the subsurface return flow and the need for region-specific models that account for local conditions such as topography, soil characteristics, and climate. Our findings provide a more accurate approach to estimating the subsurface return flow, which is crucial for improving the efficiency and sustainability of agricultural water management in Korea.

Physical Controls on Irrigation Return Flow Contributions to Stream Flow in Irrigated Alluvial Valleys

Irrigation can be a significant source of groundwater recharge in many agricultural regions, particularly in arid and semi-arid climates. Once infiltrated, irrigation recharge can travel via subsurface flowpaths that return to the river system in a lagged manner, supplementing natural streamflow weeks, months, or even years from when the irrigation was applied. In regions that experience low flows during summer and early fall, return flows can be a significant source of supplementary streamflow. Many water planning and operations models either ignore return flows or roughly approximate them with analytical solutions. Thus, return flows represent an important but often overlooked component of the hydrological exchange and overall water balance in agricultural regions. This study uses groundwater models to explore a wide range of factors that control irrigation return flow timing in irrigated alluvial valleys. A sensitivity analysis approach is used to assess how factors such as the extent of irrigated land adjacent to a stream, irrigation recharge rate, aquifer hydraulic conductivity, aquifer thickness, water table configuration, and seasonal fluctuations in stream stage control the timing of subsurface return flows. Modeling is conducted using MODFLOW models representing an irrigated alluvial valley adjacent to a stream. While a simplification of the full complexity in real systems, the models are a significant advancement from the analytical solution and provide new insight into the timescales of return flows over a broad range of possible conditions. To contextualize our modeling results, they are compared to an analytical solution commonly used for approximating return flows to evaluate its performance. Our findings show what factors and conditions influence return flow timing and control whether they contribute to streamflow over short term (months) or longer term (seasonal) time scales.

Combined influences of irrigation diversions and associated subsurface return flows on river temperature in a semi‐arid region

AbstractIrrigation activities alter water distribution and storage in arid and semi‐arid regions worldwide. The removal of water from streams can drastically impact instream flows. However, irrigation water conveyance and application onto fields can create surface and subsurface hydrologic connections, or lateral inflows, that return some of this diverted water back to streams. Prior research has shown the impact of surface water diversions from streams on downstream warming that increases stress on aquatic species. However, the combined effects of flow depletion and irrigation‐enhanced lateral inflows on stream temperature and river ecosystems remains poorly studied. To further understand these relationships, we combined intensive field monitoring over three irrigation seasons and thermal aerial imagery to identify irrigation‐enhanced subsurface lateral inflow locations and evaluate their effects on stream flow and temperature patterns over a 2.5‐km highly depleted study reach. Considering variable hydrology, weather, flow diversions, channel geometry and lateral inflows, we found irrigation‐enhanced lateral inflows were the likely explanation for buffered longitudinal and diel warming patterns that prevented stressful or lethal thermal conditions for brown trout. These localized temperature effects were more pronounced in drier years, under high diversion rates and during high solar radiation intensity. We also found that lateral inflows corresponded with greater spatial variability of stream temperatures and potential thermal refugia. Study results illustrate the potential ecological consequences of reducing irrigation‐enhanced lateral inflows and highlight the importance of hydrologic monitoring in irrigated arid river valleys. The role and preservation of these lateral inflows should be considered in water resources management related to irrigation efficiency and environmental flows.

Hydrologic modeling of reach scale fluxes from flood irrigated fields

Agricultural water is of considerable interest to water managers and policymakers as irrigation—particularly flood irrigation—accounts for the largest portion of freshwater use. However, characterization of how and when flood applied water contributes to storage and adjacent surface water bodies via return flow remains limited, particularly relative to the volume of studies addressing surface processes. This study focuses on improving our understanding of return flow paths and explicitly quantifying subsurface return flow contributions using hydro-geophysical observations coupled with a vadose zone transport flow model. Our approach, which simulates return flow in agricultural fields in a small, heavily instrumented field site in northern Wyoming, leverages daily measured values of evapotranspiration obtained from a large aperture scintillometer and atmospheric conditions, and uses seismic data to characterize heterogeneous structure of the soil profiles at field scale. We use these high-resolution, site specific datasets to parameterize a process-based vadose zone flow model that quantifies hydrologic response to flood irrigation along the stream reach, allowing us to identify return flow fluxes out of the fields. Using our approach, we found that the majority of applied irrigation water (60%) left the fields through subsurface flow paths, while only 13% was consumed by vegetation. Subsurface runoff over less-permeable soil layers, governed by the subsurface structural characteristics, was identified as the dominant subsurface flow process contributing to fluxes out of the fields. Importantly, our findings about the importance of subsurface return flow paths nuance previous research which has emphasized surface flow paths. By improving our understanding of return flow processes through a high- resolution vadose zone model that leverages climatological, geophysical, and hydrological observations, this study advances our understanding of return flow processes. Understanding how subsurface characteristics influence return flow generation in semi-arid agriculturally productive regions will help managers and policy makers balance desires for improved irrigation efficiency against concerns about irrigation-dependent ecosystem services.

Parameterization of a hydrologic model with geophysical data to simulate observed subsurface return flow paths

AbstractA portion of water not consumed by crops during flood irrigation can flow back across the surface or through the subsurface to adjacent surface water bodies and streams as return flow. Few studies have directly addressed subsurface processes governing return flow and the importance of structural complexity on hydrologic process representation. It is challenging to measure and model these subsurface flow paths using traditional hydrologic observations. In this study, we assess the impact of subsurface structural complexity on vadose zone flow representation in a two‐dimensional transport model by varying structural complexity derived from background geophysical data. We assessed four model structures each with three soil types of homogeneous hydrologic properties, two of which were evaluated with and without an anisotropy factor. Wetting front arrival times, derived from time‐lapse electrical resistivity measurements during flood irrigation field experiments, were used to evaluate the different representations of soil profile structures. These data indicated both vertical and lateral preferential flow in the subsurface during flood irrigation. Inclusion of anisotropy in the saturated hydraulic conductivity field improved the ability to model subsurface hydrologic behavior when flow processes shifted from uniform to heterogeneous flow, as occurs with lateral subsurface return flow under flood irrigation driven by a large pressure gradient. This reduced the need for detailed spatial discretization to represent these observed subsurface flow processes. The resulting simple three‐layer model structure was better able to model both the vertical and lateral flow processes than a more complex geospatial structure, suggesting that overinterpretation of smoothed inverted profiles could lead to misrepresentation of the subsurface structure.

Comparison of nutrient loss pathways: Run‐off and seepage flow in Vertisols

AbstractRecent studies have highlighted the agronomic and environmental importance of phosphorus (P) movement through the soil profile. Thus, faced with challenges such as high‐profile cases of P enrichment of surface water, better understanding of nutrient movement through soil is needed to better manage agricultural fertilizers and manures and their contribution to water quality degradation. In particular, field‐scale research is especially needed in soils with preferential flow transport pathways. Thus, we collected nitrogen (N) and P transport data in run‐off and seepage (lateral subsurface return flow) from 13 field‐ and farm‐scale watersheds on Vertisols in Central Texas for a 14‐year period. For 2004–2017, seepage accounted for ~20% of the total surface flow, and nutrient concentrations were generally similar in run‐off and seepage. As surface run‐off contributed ~80% of the flow, it follows that median annual N and P loads in run‐off were significantly greater than in seepage for every watershed. N loads in both run‐off and seepage flow from cultivated land were an order of magnitude greater than in native prairie and improved pasture, and the highest run‐off and seepage P loads both occurred on cultivated land with organic fertilizer sources. Increasing watershed scale (size) did not to produce consistent patterns in N or P loss in run‐off or seepage. Land use and watershed scale produced significant differences in seepage volume but did not affect run‐off volumes or total surface flow/rainfall. Although less significant in terms of total offsite flux, nutrient movement in vadose zones has important agronomic and environmental implications as considerable N and P are transported through and within the root zone and eventually offsite. And in terms of P, this contradicts the traditionally held scientific viewpoint that P movement through the vadose zone is unimportant agronomically and environmentally.

Return flows from beaver ponds enhance floodplain-to-river metals exchange in alluvial mountain catchments

River to floodplain hydrologic connectivity is strongly enhanced by beaver- (Castor canadensis) engineered channel water diversions. The hydroecological impacts are wide ranging and generally positive, however, the hydrogeochemical characteristics of beaver-induced flowpaths have not been thoroughly examined. Using a suite of complementary ground- and drone-based heat tracing and remote sensing methodology we characterized the physical template of beaver-induced floodplain exchange for two alluvial mountain streams near Crested Butte, Colorado, USA. A flowpath-oriented perspective to water quality sampling allowed characterization of the chemical evolution of channel water diverted through floodplain beaver ponds and ultimately back to the channel in ‘beaver pond return flows’. Subsurface return flow seepages were universally suboxic, while ponds and surface return flows showed a range of oxygen concentration due to in-situ photosynthesis and atmospheric mixing. Median concentrations of reduced metals: manganese (Mn), iron (Fe), aluminum (Al), and arsenic (As) were substantially higher along beaver-induced flowpaths than in geologically controlled seepages and upstream main channel locations. The areal footprint of reduced return seepage flowpaths were imaged with surface electromagnetic methods, indicating extensive zones of high-conductivity shallow groundwater flowing back toward the main channels and emerging at relatively warm bank seepage zones observed with infrared. Multiple-depth redox dynamics within one focused seepage zone showed coupled variation over time, likely driven by observed changes in seepage rate that may be controlled by pond stage. High-resolution times series of dissolved Mn and Fe collected downstream of the beaver-impacted reaches demonstrated seasonal dynamics in mixed river metal concentrations. Al time series concentrations showed proportional change to Fe at the smaller stream location, indicating chemically reduced flowpaths were sourcing Al to the channel. Overall our results indicated beaver-induced floodplain exchanges create important, and perhaps dominant, transport pathways for floodplain metals by expanding chemically-reduced zones paired with strong advective exchange.

Uniform and lateral preferential flows under flood irrigation at field scale

AbstractFlood irrigation is globally one of the most used irrigation methods. Typically, not all water that is applied during flood irrigation is consumed by plants or lost to evaporation. Return flow, the portion of applied water from flood irrigation that returns back to streams either via surface or subsurface flow, can constitute a large part of the water balance. Few studies have addressed the connection between vertical and lateral subsurface flows and its potential role in determining return flow pathways due to the difficulty in observing and quantifying these processes at plot or field scale. We employed a novel approach, combining induced polarization, time‐lapse electrical resistivity tomography, and time‐lapse borehole nuclear magnetic resonance, to identify flow paths and quantify changes in soil hydrological conditions under nonuniform application of flood irrigation water. We developed and tested a new method to track the wetting front in the subsurface using the full range of inverted resistivity values. Antecedent soil moisture conditions did not play an important role in preferential flow path activation. More importantly, boundaries between lithological zones in the soil profile were observed to control preferential flow pathways with subsurface run‐off occurring at these boundaries when saturation occurred. Using the new method to analyse time‐lapse resistivity measurements, we were able to track the wetting front and identify subsurface flow paths. Both uniform infiltration and preferential lateral flows were observed. Combining three geophysical methods, we documented the influence of lithology on subsurface flow processes. This study highlights the importance of characterizing the subsurface when the objective is to identify and quantify subsurface return flow pathways under flood irrigation.

USDA‐ARS Riesel Watersheds, Riesel, Texas, USA: Water quality research database

AbstractThe 75 year legacy database including discharge, sediment loss, land management, and meteorological data for the USDA‐ARS Riesel Watersheds, Riesel, TX, USA has been available on the web for more than a decade (www.ars.usda.gov/spa/hydro‐data) and used in numerous studies and publications; however, only recently have these data been added to the Sustaining the Earth's Watersheds, Agricultural Research Data System (STEWARDS) database (www.nrrig.mwa.ars.usda.gov/stewards/stewards.html). In addition, water quality data including dissolved inorganic N and P compounds measured from more than 1000 storm runoff events, 1300 base flow sampling events (lateral subsurface return flow or seepage flow), and 157 precipitation events through 2012 were added. The objectives of this manuscript are to present relevant background information on these data, summarize the data collection and analysis methodology, present the measured data along with cursory analyses, and convey the commitment of the USDA‐ARS Riesel Watersheds to long‐term data accessibility and database enhancement for water quality data and research.

Modeling irrigation return flow for the return flow reuse system in paddy fields

This research is to construct a water balance model to estimate the amount of return flow in an irrigation system. A simple computation framework for the model was established to include various irrigation applications in cropping seasons. The model was able to estimate evapotranspiration, deep percolation into groundwater aquifer, and return flow. Return flow can be split into two parts, which are surface and subsurface return flows. The water balance model was then applied at the irrigation system (rotational block No. 11-2 of five paddy field units) which is operated by the Taoyuan Irrigation Association in Taiwan as an example. Two study cases were simulated, in which one was for using return flow and the other one was for using no return flow. The study period for the model simulations is the first rice cropping term in 2010 which was from February 16 to July 10. As a result, return flows calculated by the model were 27, 27, 34, and 39% of outflows for sandy loam, sandy clay, clay loam, and light clay soil, respectively. Irrigation water at the downstream field unit with use of return flow was supplemented by the upstream field units, and the amount is 5–8% of irrigation water for using no return flow. Furthermore, it can be seen from the simulations that increases in irrigation water provide increases of return flow. Increases of irrigation water result in slight increases of subsurface return flow, while increases of irrigation water cause nearly none of change in deep percolation.

A new algorithm to calculate weighted flow-accumulation from a DEM by considering surface and underground stormwater infrastructure

This study presents a new algorithm, called the Weighted Adaptive Stormwater Infrastructure (WASI) algorithm, to calculate weighted flow-accumulation of surface runoff by considering man-made stormwater infrastructure that modifies natural drainage flow paths across the landscape. The WASI algorithm incorporates ancillary data sets required to determine the depth of effective rainfall (i.e., rate of precipitation plus subsurface return flow, rate of evapotranspiration plus infiltration, duration of rainfall, depth of initial absorption to ground) and a priori information on surface and underground stormwater infrastructure into grid-based hydrologic analysis using a DEM. Using new software, named the WASI surface runoff calculator (implemented as part of this study), the WASI algorithm can be applied to urban and suburban areas where both surface and underground stormwater infrastructure is installed. This paper describes the concept and details of the WASI algorithm, as well as its applications to synthetic and real-world data sets which revealed that the new software can provide more accurate hydrologic information than other commercial GIS software such as IDRISI and ArcGIS.

How do disconnected macropores in sloping soils facilitate preferential flow?

AbstractRunoff from hillslopes is generated by processes such as infiltration‐excess and saturation overland flow, subsurface stormflow and subsurface return flow. Preferential flow through macropores can affect any one of these runoff generation processes. Field studies at the Hitachi Ohta Experimental Watershed in Japan have noted that self‐organization processes may manifest the connectivity of such subsurface flow paths, particularly via macropores, from hillslopes to stream channels. It is well established from soil physics principles that the connectivity of macropore networks depends on soil wetness and this has been shown experimentally at Hitachi Ohta where subsurface flow and streamflow respond to thresholds of wetness. Numerical solutions to the three‐dimensional Richards equation are derived for a sloping soil block containing a population of disconnected macropores of various sizes, shapes and orientations. Solutions for the case of steady water flux applied to the surface of the soil block are evaluated to determine the conditions where the disconnected macropores become active in the flow process. Results show that subsurface flow is directed through the preferential flow network in the saturated portion of the soil but bypasses the macropores in the drier regions. The preferential flow network expands as the degree of saturation increases. The expanding network of active macropores leads to less resistance to overall flow in the domain and access to increased volumes of the flow domain. Although the individual macropores are disconnected, it is argued that large localized hydraulic gradients can potentially lead to preferred zones of subsurface erosion. In addition to the importance of these findings related to stormflow generation in catchments, they add support to the concept of self‐organization of subsurface flow systems in soils. Copyright © 2010 John Wiley & Sons, Ltd.

Langmuir circulation driving sediment entrainment into newly formed ice: Tank experiment results with application to nature (Lake Hattie, United States; Kara Sea, Siberia)

Langmuir circulation (Lc) was generated under freezing conditions in saltwater tank experiments through surface wind stress and cross‐waves interacting with subsurface return flow. Fine‐grained sediments distributed in the tank prior to frazil crystal formation were aligned in parallel streaks in Lc bottom convergence zones. Downwelling at Lc surface convergence zones aligned floating frazil in wind‐parallel rows, and individual crystals rotated on helical paths down to the tank bottom and up again to the surface. The crystals interacted with suspended particles in the water column, and with sediment on the tank bottom, preferentially collecting fine‐grained particles and enhancing their entrainment into new ice. Evidence includes higher sediment concentrations in ice and ice‐interstitial water (ice pore water) as compared to the tank water. Both tank ice and ice interstitial water contain more silt‐sized particles than tank water suspension load and tank bottom sediment. Sand is reduced in the ice, and clay is about the same concentration in all samples. This points to preferential entrainment of fine particles in newly formed ice supported by Lc‐driven circulation. Comparable results of Lc‐supported ice particle entrainment were found in Lake Hattie. Comparison of ice sediment from tank experiments run with Kara Sea material to ice particles from the natural Kara setting showed both types of ice sediment have very similar grain size distributions and mineralogical compositions. Results from experiments and nature help to better understand the potentially Lc‐driven entrainment of sediment into ice formed in shallow freezing waters.

Runoff and soil loss relationships for the Texas Blackland Prairies ecoregion

Hydrological and soil loss data have been collected since 1937 at the USDA-ARS Grassland Soil and Water Research Laboratory near Riesel, TX. Data from the site, originally named the Blacklands Experimental Watershed, have played a vital role in the evaluation of conservation management practices to limit soil erosion and offsite herbicide transport and in the development of several watershed models used worldwide. The entire record of precipitation, runoff, sediment loss, management practices, and limited meteorological information is publicly available (http://www.ars.usda.gov/spa/hydro-data). The data represent a valuable regional resource for use in water supply modeling, rural land development, and agricultural land management. Results of the present analyses confirmed the importance of soil-water phases to temporal runoff patterns in the Texas Blackland Prairies ecoregion. Little runoff occurs in the dry soil-water phase (avg. = 2-9 mm/month), but substantial surface runoff and lateral subsurface return flow occurs in the saturated phase (avg. = 19-28 mm/month). Strong linear relationships (P< 0.0001) were determined between watershed size and annual peak flow rates for return intervals from 2 to 100 yr. Long-term data indicate a drastic reduction in soil loss from small grain production compared to row crop production due to the presence of soil cover in both the spring and fall high precipitation periods. Thus, utilization of a winter cover crop in row crop production or conversion from row crop to small grain production can be effective in reducing offsite transport of sediment and associated contaminants, which may be important in watersheds with substantial agricultural contribution to water quality impairment. Published by Elsevier B.V.

Impact of Irrigation Cessation on Wetland Communities within the Elk Ranch, Grand Teton National Park, Moose, Wyoming

Riparian ecology research in Arizona and California has documented the likelihood of a subsurface linkage between irrigation, especially flood-irrigation, and riparian function (Smith et al. 1989; Stromberg et al. 1996). Initial groundwater monitoring results from rural New Mexico indicate water tables rose 1 to 2m after the onset of field irrigation and subsurface flow paths towards the Rio Grande River developed soon after (Fernald et al. 2008). Results from a study of wetlands in southeastern Wyoming suggest that declining flood-irrigation levels would lead to a reduction in the total area of wetlands and related areas of wetland vegetation types in the Laramie Basin (Peck and Lovvorm 2001). Stringham et al. (1998) have reported further evidence for a linkage between irrigation and riparian function. These Oregon researchers noted lower water temperatures in stream reaches receiving subsurface return flows from irrigated hayfields than similar reaches flowing through non-irrigated lands. This information is timely because Grand Teton National Park (GTNP) managers have begun an evaluation of historic irrigation operations within the Park and are endeavoring to learn how cessation of flood irrigation will affect Park wetlands. The historically irrigated hayfields at the Elk Ranch provide an opportunity to address the Park Service's informational needs through identification of vegetation composition, soil physical characteristics and groundwater patterns associated with irrigated and naturally occurring wetlands. Successful description of patterns unique to natural wetlands will provide an avenue for predicting which Park wetlands would remain functional should irrigation efforts be brought to a close. Development of criteria for identifying naturally occurring wetlands could also serve as a basis for identifying areas for wetland mitigation and rehabilitation elsewhere in GTNP and the mountain valleys of the Northern Rocky Mountains.

Agroecological Impacts from Salinization and Waterlogging in an Irrigated River Valley

Extensive field data and calibrated flow and salt-transport models characterize the spatial and temporal patterns of salinity and waterlogging in an irrigated western river valley. Over three irrigation seasons, average seasonal aquifer recharge from irrigated fields in a 50,600 ha study area ranges from 0.59 to 0.99 m , including contribution from precipitation. The salinity of irrigation water varies from 618 to 1,090 mg∕L . The water table is shallow, with 16 to 33% of irrigated land underlaid by an average water table less than 2 m deep. Average water table salinity ranges from 2,680 to 3,015 mg∕L , and average soil salinity from 2,490 to 3,860 mg∕L . Crop yield reductions from salinity and waterlogging range from 0 to 89% on fields, with regional averages ranging from 11 to 19%. Annual salt loading to the river from subsurface return flows, generated in large part by dissolution from irrigation recharge, averages about 533 kg∕irrigated ha per km . Upflux from shallow water tables under fallow ground ...

SUBSURFACE RETURN FLOW AND GROUND WATER RECHARGE OF TERRACE FIELDS IN NORTHERN TAIWAN1

ABSTRACT: This study estimates subsurface return flow and effective ground water recharge in terraced fields in northern Taiwan. Specifically, a three dimensional model, FEMWATER, was applied to simulate percolation and lateral seepage in the terraced fields under various conditions. In the terraced paddy fields, percolation mainly moves vertically downward in the central area, while lateral seepage is mainly focused around the bund. Although the simulated lateral seepage rate through the bund exceeded the percolation rate in the central area of the paddy field, annual subsurface return flow at Pei‐Chi and Shin‐Men was 0.17 × 106 m3 and 0.37 × 106 m3, representing only 0.17 percent and 0.21 percent of the total irrigation water required for rice growth at Pei‐Chi and Shin‐Men, respectively. For upland fields, the effective ground water recharge rate during the second crop period (July to November) exceeded that during the first crop period (January to May) because of the wet season in the second crop period. Terraced paddy fields have the most efficient ground water recharge, with 21.2 to 23.4 percent of irrigation water recharging to ground water, whereas upland fields with a plow layer have the least efficient ground water recharge, with only 4.8 to 6.6 percent of irrigation water recharging to ground water. The simulation results clearly revealed that a substantial amount of irrigation water recharges to ground water in the terraced paddy, while only a small amount of subsurface return flow seeps from the upstream to the downstream terraced paddy. The amounts of subsurface flow and ground water recharge determined in the study are useful for the irrigation water planning and management and provide a scientific basis to reevaluate water resources management in the terrace region under irrigated rice.

Analysis of percolation and seepage through paddy bunds

This study investigates percolation and seepage through the bunds of flat and terraced paddies. Field experiments were conducted in Hsin-Pu of Hsin-Chu County, Taiwan, to measure the soil water content of various types of bund. Measurements revealed that the soil was unsaturated along the sloped surface of the terrace. Experimental results also indicated that seepage face flow did not develop even after 2 days of heavy rainfall. A three-dimensional model, FEMWATER, was adopted to simulate percolation and lateral seepage under various bund conditions. In a flat paddy, the rate of percolation of bunds under which a plow sole was located, was 0.40 cm d −1, close to the average infiltration rate of a flooded paddy. The percolation of the bund without plow sole was 0.85 cm d −1, or double the average infiltration rate of a flooded paddy. Infiltration in the central area of a terraced paddy is mainly vertically downward, whereas flow near the bund is predominantly lateral. The paddy field near the bund has a high hydraulic gradient. The simulated infiltration flux into the bund (1.47 cm d −1) after 85 days of rice cultivation exceeded that into the central area (0.54 cm d −1) by a factor of 2.72. The final percolation flux from the bund (1.24 cm d −1) also exceeded the final percolation from the plow sole (0.68 cm d −1) by a factor of 1.82. The lateral seepage fluxes through the bund, downward and upward along the slope surface, are 2.01 and −2.12 cm d −1, respectively. However, the lateral seepage flux does not fully saturate the surface of the hillside soil. A simulation clearly shows that the seepage upstream of the paddy field does not move water downstream and is reused as subsurface return flow. Both experimental and simulation results clarify the mechanisms of water movement in the terraced paddy and reveal the existence of an unsaturated seepage face along the sloping surface of the terraced field.

Application of a geographic information system for conceptual rainfall–runoff modeling

Geographic information systems (GIS) offer many new opportunities for hydrological modeling. They can be used to form spatially distributed models of watershed. However, some problems of this approach, e.g. the parameterization of physically based models, are not resolved yet. Conceptual models of the meso-scale still have a great practical importance. In this paper one approach is presented: how statistical descriptions of distributed catchment characteristics could be used to consider spatial heterogeneity within conceptual models. Three semi-distributed modules are presented. The three components are combined to a hydrological model including feedback components between surface flow and infiltration and between subsurface return flow and surface flow in saturated areas. The model was set up to use spatially distributed information about catchment characteristics for the estimation of its parameters. By a direct estimation of some model parameters from a GIS-based analysis of the catchment characteristics, the number of calibration parameters can be reduced. In the second part it is shown how the application of this model to different catchments within a region can benefit from boundary conditions for optimization, which are derived from a GIS considering the differences of catchment characteristics.

A mathematical concept of runoff prediction model for small tropical catchment areas

After the Horton overland flow concept became legend until the early 1960s, new findings then revised this concept for implementation in several catchment area conditions. Runoff processes in tropical regions seem to have specific characteristics such as thick humus on the top layer of the catchment area and non-uniformity of hydraulic conductivity vertically and horizontally across the stream. To overcome the runoff problems in tropical regions, a reasonable model needed to be developed to predict generated runoff. Based on the concept that saturation condition in a catchment generates non-uniformly during the course of a rainfall, and that the storage capability of top layer in intercepting rain-water leads to subsurface and return flow, the water balance in a segment of the catchment area is calculated. The mechanism of flow in the model is that subsurface flow and return flow, which are a function of hydraulic conductivity, and surface runoff on a segment are integrated for the whole areas. The generation of flow using the model seems to indicate that runoff peaks will commence after each segment is saturated starting from the valley bottom. The lag time for initial flow to alter significantly depends on the data on stream-flow before the storm began that indicate the soil moisture and depth of the temporary water table. Generally, though, the model is very flexible, so that data on each segment subject to the wide variety of physical conditions of the catchment area characteristic of tropical regions can easily be implemented.