Soil cracking effects on hydrological and erosive processes: a study case in Mediterranean cultivated vertisols

Shrink–swell soils, such as those in a Mediterranean climate regime, can cause changes in terms of hydrological and erosive responses because of the changing soil water storage conditions. Only a limited number of long‐term studies have focused on the impacts on both hydrological and erosive responses and their interactions in an agricultural environment. In this context, this study aims to document the dynamics of cracks, runoff and soil erosion within a small Mediterranean cultivated catchment and to quantify the influence of crack processes on the water and sediment supplied to a reservoir located at the catchment outlet using water and sediment measurements at a cultivated field outlet as baseline. Detailed monitoring of the presence of topsoil cracks was conducted within the Kamech catchment (ORE OMERE, Tunisia), and runoff and suspended sediment loads were continuously measured over a long period of time (2005–2012) at the outlets of a field (1.32 ha) and a catchment (263 ha). Analysis of the data showed that topsoil cracks were open approximately half of the year and that the rainfall regime and water table level conditions locally control the seasonal cracking dynamics. Topsoil cracks appeared to seriously affect the generation of runoff and sediment concentrations and, consequently, sediment yields, with similar dynamics observed at the field and catchment outlets. A similar time lag in the seasonality between water and sediment delivery was observed at these two scales: although the runoff rates were globally low during the presence of topsoil cracks, most sediment transport occurred during this period associated with very high sediment concentrations. This study underlines the importance of a good prediction of runoff during the presence of cracks for reservoir siltation considerations. In this context, the prediction of cracking effects on runoff and soil erosion is a key factor for the development of effective soil and water management strategies and downstream reservoir preservation. Copyright © 2016 John Wiley & Sons, Ltd.     

High-frequency monitoring of hydrological and biogeochemical fluxes in forested catchments of southern Chile

The variability of rainfall-dependent streamflow at catchment scale modulates many ecosystem processes in wet temperate forests. Runoff in small mountain catchments is characterized by a quick response to rainfall pulses which affects biogeochemical fluxes to all downstream systems. In wet-temperate climates, water erosion is the most important natural factor driving downstream soil and nutrient losses from upland ecosystems. Most hydrochemical studies have focused on water flux measurements at hourly scales, along with weekly or monthly samples for water chemistry. Here, we assessed how water and element flows from broad-leaved, evergreen forested catchments in southwestern South America, are influenced by different successional stages, quantifying runoff, sediment transport and nutrient fluxes during hourly rainfall events of different intensities. Hydrograph comparisons among different successional stages indicated that forested catchments differed in their responses to high intensity rainfall, with greater runoff in areas covered by secondary forests (SF), compared to old-growth forest cover (OG) and dense scrub vegetation (CH). Further, throughfall water was greatly nutrient enriched for all forest types. Suspended sediment loads varied between successional stages. SF catchments exported 455 kg of sediments per ha, followed by OG with 91 kg/ha and CH with 14 kg/ha, corresponding to 11 rainfall events measured from December 2013 to April 2014. Total nitrogen (TN) and phosphorus (TP) concentrations in stream water also varied with rainfall intensity. In seven rainfall events sampled during the study period, CH catchments exported less nutrients (46 kg/ha TN and 7 kg/ha TP) than SF catchments (718 kg/ha TN and 107 kg/ha TP), while OG catchments exported intermediate sediment loads (201 kg/ha TN and 23 kg/ha TP). Further, we found significant effects of successional stage attributes (vegetation structure and soil physical properties) and catchment morphometry on runoff and sediment concentrations, and greater nutrients retention in OG and CH catchments. We conclude that in these southern hemisphere, broad-leaved evergreen temperate forests, hydrological processes are driven by multiple interacting phenomena, including climate, vegetation, soils, topography, and disturbance history.     

Runoff and peak flow responses to timber harvest and forest age in southern Chile

Runoff and peak flows in three experimental catchments with different forest conditions were analysed in a rainy temperate climate in southern Chile. The hydrological effects of clearcutting a Pinus radiata plantation covering 79·4% of the La Reina catchment were studied by analysing runoff and peak flows in the pre‐ and post‐harvesting periods. Mean annual runoff increased 110% after timber harvesting. Clearcutting generated a 32% mean increase in peak flows. Peak flow and runoff were examined in two adjacent catchments with different forest conditions. The older plantation in Los Ulmos 1 increasingly consumed more water than the younger plantation established at Los Ulmos 2, whereas differences in peak flows between these two catchments were not significant. Runoff and peak flows comparisons not only reflected changes in forest cover, but also the effect of rainfall characteristics during the study periods and the basins' morphologies. Comparisons between pre‐ and post‐harvesting peak discharges must be undertaken with caution, because a previous analysis at La Reina using a partial set of data overestimated changes in peak flows after timber harvesting. Copyright © 2005 John Wiley & Sons, Ltd.     

Increasing annual runoff—broadleaf or coniferous forests?

According to widely held belief, annual evapotranspiration (ET) for broadleaf forests is less than that for coniferous forests, resulting in higher annual runoff for broadleaf forests. We processed 82 catchment runoff and 126 interception loss data from temperate regions and found that although the belief is valid under conditions of broadleaf deciduous forests and high winter precipitation (e.g. the United States), it is invalid under conditions of broadleaf evergreen forests (e.g. New Zealand) or low winter precipitation (e.g. Japan). Thus, forest management policies based on this belief should be reconsidered on the basis of our results for regions with broadleaf evergreen forests or low winter precipitation. Copyright © 2010 John Wiley & Sons, Ltd.     

Temporal variability in hydrological response within a small catchment with badland areas,central Pyrenees

Abstract

The lower Araguás catchment, central Pyrenees, is characterized by extensive badlands (25% of the total catchment), whereas the upper catchment is covered by dense plantation forest. The catchment (45 ha) has been monitored since October 2005 with the aim of studying its hydrological response. The 44 floods recorded over this period were analysed to identify the factors that control the rainfall—runoff relationship. The first relevant feature of the catchment was its responsiveness. The catchment reacted to all rainfall events, but the irregular nature of the hydrological response was the most characteristic feature of the response. No single variable could explain the response of the Araguás catchment. It was found that stormflow coefficients mainly depend on the combination of rainfall volume and antecedent baseflow. A significant correlation was observed between maximum rainfall intensity and peak flow values. The shapes of the different hydrographs are very similar, regardless of the peak flow magnitude; they show a short time lag, relatively narrow peak flow, and steep recession limb. This indicates a large contribution by overland flow, resulting mainly from the generation of infiltration excess runoff in badland areas.     

Exploring suspended sediment delivery dynamics of two Mediterranean nested catchments

This paper investigates suspended sediment transport and dynamics of two nested agricultural lowland Mediterranean catchments with a difference of two orders of magnitude in the surface area (i.e., 1 and 264 km²). The effects of the drainage catchment area over the specific suspended sediment yield are assessed by using the nested approach over various timeframes. A detailed analysis of the rainfall–runoff–sediment transport relationships during the 2‐year study period shows that the hydrological and sedimentological responses were extremely variable for both catchments. Very low or no correlations were observed between the rainfall intensity and the selected hydrological variables and sediment loads. However, remarkable or high correlations were obtained between the rainfall intensity and the maximum and average suspended sediment concentrations, indicating that rainfall per unit time has little control on the hydrological response, but that, simultaneously, its high‐erosive power triggers sediment production, increasing the sedimentary response of the catchments. This study also illustrates how sediment is mainly transported during floods, producing predominantly clockwise hysteretic loops. Moreover, the small headwater catchment exerts a reduced (or even negligible) effect over the hydro‐sedimentary response of the larger downstream catchment, caused by the reduced sediment availability in a landscape with an inherent disconnection of the sediment pathways.     

Summer flows in experimental catchments with different forest covers, Chile

Runoff and peak flows in four experimental catchments with different land uses are analyzed for summer periods. The catchments have a rainy temperate climate with annual precipitations between 2000 and 2500 mm, 70% of which is concentrated in the winter period between May and August. The final harvest of the forest plantation in one of these catchments generated increases in summer runoff. Also, differences between the maximum instantaneous discharge and the flow at the beginning of the storm then almost duplicated those registered in rainfall events of similar magnitude when the catchment was fully forested. Runoff analysis in this catchment is difficult because the two post-harvesting summer periods are much wetter than the two pre-harvesting ones but a double mass analysis shows the effect of harvesting clearly. In a paired catchment study, low cover in one of the two neighbour catchments explains higher direct runoff and base flows although lower maximum instantaneous specific discharge occurred in the less vegetated but larger catchment. Low vegetation cover explains increases in summer flows, although the size, topography, rainfall conditions, road density, extent of affected area and runoff generation processes play an important role in the hydrological effects of different land uses.     

Impacts of wildfire and post-fire land management on hydrological and sediment processes in a humid Mediterranean headwater catchment

The extensive afforestation of the Mediterranean rim of Europe in recent decades has increased the number of wildfire disturbances on hydrological and sediment processes, but the impacts on headwater catchments is still poorly understood, especially when compared with the previous agricultural landscape. This work monitored an agroforestry catchment in the north-western Iberian Peninsula, with plantation forests mixed with traditional agriculture using soil conservation practices, for one year before the fire and for three years afterwards, during which period the burnt area was ploughed and reforested. During this period, continuous data was collected for meteorology, streamflow and sediment concentration at the outlet, erosion features were mapped and measured after major rainfall events, and channel sediment dynamics were monitored downstream from the agricultural and the burnt forest area. Data from 202 rainfall events with over 10 mm was analysed in detail. Results show that the fire led to a notable impact on sediment processes during the first two post-fire years, but not on streamflow processes; this despite the small size of the burnt area (10% of the catchment) and the occurrence of a severe drought in the first year after the fire. During this period, soil loss at the burnt forest slopes was much larger than that at most traditionally managed fields, and, ultimately, led to sediment exhaustion. At the catchment scale, storm characteristics were the dominant factor behind streamflow and sediment yield both before and after the fire. However, the data indicated a shift from detachment-limited sediment yield before the fire, to transport-limited sediment yield afterwards, with important increases in streamflow sediment concentration. This indicates that even small fires can temporarily change sediment processes in agroforestry catchments, with potential negative consequences for downstream water quality.     

Comparative analysis of the response of various land covers to an exceptional rainfall event in the central Spanish Pyrenees,October 2012

The hydrological and geomorphological effects of an exceptional rainstorm event that occurred in the central Spanish Pyrenees during 19–21 October 2012 were studied in five experimental catchments under various land covers: (i) subhumid badlands; (ii) dense forest; (iii) an abandoned farmland area recolonized by shrubs and forest patches; and (iv) subalpine grasslands. Hydrographs and sedigraphs demonstrated that vegetation cover is a major factor affecting the control of floods even during exceptional rainstorms, at least at the spatial scale at which the phenomenon was studied (catchment sizes: 0.3–2.8 km2) and under dry catchment conditions. The combined precipitation over the two days (c. 250 mm) was the greatest for any two‐day event recorded since 1950 in the central‐western Pyrenees for all but one of the stations in the study. Five pulses of most intense rainfall were recorded. The forested catchment did not react to the two most intense rainfall pulses, because of the very low antecedent level of the water table. The main peak flow occurred only when at least a part of the catchment was saturated. The abandoned farmland catchment had two small peak discharges at the beginning of the event, which were produced by infiltration excess overland flow from eroded areas close to the main stream. During the third most intense rainfall period a large part of this catchment contributed to runoff and a relatively high peak discharge was produced. The badland catchment reacted immediately from the beginning of the rainstorm, yielding very high discharges accompanied by high suspended sediment concentrations. The subalpine catchment showed a hydrograph mirroring the hyetograph, with brief but intense hydrological responses to increased precipitation, because of the marked gradients and the presence of bare rock in the headwaters. A high volume of bedload was carried during the peak discharge. Copyright © 2013 John Wiley & Sons, Ltd.     

Effect of forest on annual water yield in the mountains of an arid inland river basin: a case study in the Pailugou catchment on northwestern China's Qilian Mountains

The effect of forests on annual water yield is an unresolved central issue in forest hydrology despite years of study. There has been a particular shortage of research in the mountains of arid inland river basins. In the present study, we examined the effects of forests on hydrology using data on precipitation, evaporation, canopy interception, transpiration, and runoff from 1994 to 2008 for the Pailugou catchment of northwestern China's Qilian Mountains. We modelled the water balance to assess the contribution of different vegetation types to annual water yield. In our study area, Picea crassifolia forest covered 38·5% of the catchment area, but contributed little to annual water yield. For an annual average precipitation of 407·1 mm (from 2003 to 2008) at an elevation of 2700 m, the runoff depth from the forest was 11·6 mm, accounting for only 3·5% of total annual water yield of the catchment. For an annual average precipitation of 374·1 mm (from 1994 to 2002), the runoff depth from the forest was ? 14·3 mm (i.e. 5·9% of total annual water yield of the catchment was consumed to sustain tree growth). This has significant implications, because forests are increasingly being planted in the Qilian Mountains, and this may decrease the downstream water supply. Thus, the relationship between the forest and water yield must be better understood to permit the establishment of an appropriate regional level of forest cover. Copyright © 2011 John Wiley & Sons, Ltd.     

Evaluating and modelling the hydrological and erosive behaviour of an olive orchard microcatchment under no‐tillage with bare soil in Spain

Despite the high risk of erosion in olive orchards located in mountainous areas in Spain, little research has been carried out to account for the complexity and interaction of the natural processes of runoff and soil erosion on the catchment scale or small catchment scale. In this study, a microcatchment of 6·7 ha in a mountainous area under no‐tillage farming with bare soil was set up to record runoff and sediment. Soil erosion and runoff patterns were monitored over a two‐year period. Totally, 22 events were observed. The data were analysed, and then used to calibrate the AnnAGNPS model, which allowed us to complete the data period and describe the hydrological and erosive behaviour on a monthly and annual basis. A high variability in catchment responses was observed, due to differences in the storms and to the effect of the surface soil moisture content. Maximum intensities of 10 and 30 min determined the final runoff values while the total sediment loads were dependent on the rainfall depth. The impact of management on the reduction of porosity can explain the relationship between runoff and intensity in the microcatchment. However, the impact of the spatial scale meant that the transport of sediment required substantial rainfall depths to ensure a continuous flow from the hillslopes. The results of the calibration (E >0·60 and r >0·75) on the event and monthly scale confirmed the applicability of AnnAGNPS to predict runoff and erosion in the microcatchment. The predicted average runoff coefficient was 3·3% for the study period and the total average sediment loads, 1·3 Mg/ha/yr. Despite these low values, the model simulation showed that much larger runoff coefficients and soil losses can be expected for periods with several consecutive years in which the annual rainfall depth was over 500 mm. The use of cover is recommended to prevent the high levels of erosion associated with these conditions. Copyright © 2009 John Wiley & Sons, Ltd.     

Estimating the sediment trap efficiency of intermittently dry reservoirs: lessons from the Kruger National Park,South Africa

The assessment of sediment yield from reservoir siltation requires knowledge of the reservoir's sediment trap efficiency (TE). Widely used approaches for the estimation of the long‐term mean TE rely on the ratio of the reservoir's storage capacity (C) to its catchment size (A) or mean annual inflow (I). These approaches have been developed from a limited number of reservoirs (N ≤ 40), most of them located in temperate climate regions. Their general applicability to reservoirs receiving highly variable runoff such as in semi‐arid areas has been questioned. Here, we examine the effect of ephemeral inflow on the TE of 10 small (≤ 280 × 10³ m³), intermittently dry reservoirs located in the Kruger National Park. Fieldwork was conducted to determine the storage capacity of the reservoir basins. The frequency and magnitude of spillage events was simulated with the daily time step Pitman rainfall–runoff model. Different runoff scenarios were established to cope with uncertainties arising from the lack of runoff records and imperfect input data. Scenarios for the relationship between water and sediment discharge were created based on sediment rating curves. Taking into account uncertainties in hydrological modelling, uncertainties of mean TE estimates, calculated from all scenarios (N = 9), are moderate, ranging from ±6 to ±11% at the 95% confidence level. By comparison, estimating TE from the storage capacity to catchment area (C/A) ratio induces high uncertainty (ranges of 35 to 65%), but this uncertainty can be confined (15 to 33%) when the latter approach is combined with hydrological modelling. Established methods relying on the storage capacity to mean annual inflow (C/I) ratio most probably lead to an overestimation of the TE for the investigated reservoirs. The approach presented here may be used instead to estimate the TE of small, intermittently dry reservoirs in semi‐arid climate regions. Copyright © 2017 John Wiley & Sons, Ltd.     

Brixenbach research catchment: Quantification of runoff process proportions in a small Alpine catchment depending on soil moisture states and precipitation characteristics

The Brixenbach valley is a small Alpine torrent catchment (9.2 km², 820–1950 m a.s.l., 47.45°, 12.26°) in Tyrol, Austria. Intensive hydrological research in the catchment since more than 12 years, including a hydrogeological survey, pedological and land use mapping, measurements of precipitation, runoff, soil moisture and infiltration as well as the conduction of rainfall simulations, has contributed to understand the hydrological response of the catchment, its subcatchments and specific sites. The paper presents a synthesis of the research in form of runoff process maps for different soil moisture states and precipitation characteristics, derived with the aid of a newly developed Soil-hydrological model. These maps clearly visualize the differing runoff reaction of different subcatchments. The pasture dominated areas produce high surface flow rates during short precipitation events (1 h, 86 mm) with high rainfall intensity, whilst the forested areas often develop shallow subsurface flow. Dry preconditions lead to a slight reduction of surface flow, long rainfall events (24 h, 170 mm) to a dominance of deep subsurface flow and percolation.     

An analysis of the impact of spatial variability in rainfall on runoff and sediment predictions from a distributed model

This paper investigates the effect of introducing spatially varying rainfall fields to a hydrological model simulating runoff and erosion. Pairs of model simulations were run using either spatially uniform (i.e. spatially averaged) or spatially varying rainfall fields on a 500‐m grid. The hydrological model used was a simplified version of Thales which enabled runoff generation processes to be isolated from hillslope averaging processes. Both saturation excess and infiltration excess generation mechanisms were considered, as simplifications of actual hillslope processes. A 5‐year average recurrence interval synthetic rainfall event typical of temperate climates (Melbourne, Australia) was used. The erosion model was based on the WEPP interrill equation, modified to allow nonlinear terms relating the erosion rate to rainfall or runoff‐squared. The model results were extracted at different scales to investigate whether the effects of spatially varying rainfall were scale dependent. A series of statistical metrics were developed to assess the variability due to introducing the spatially varying rainfall field. At the catchment (approximately 150 km²) scale, it was found that particularly for saturation excess runoff, model predictions of runoff were insensitive to the spatial resolution of the rainfall data. Generally, erosion processes at smaller sub‐catchment scales, particularly when the sediment generation equation had non linearity, were more sensitive to spatial rainfall variability. Introducing runon infiltration reduced the total runoff and sediment yield at all scales, and this process was also most sensitive to the rainfall resolution. Copyright © 2011 John Wiley & Sons, Ltd.     

Comparative application of two mathematical models to predict sedimentation in Yermasoyia Reservoir,Cyprus

The Yermasoyia Reservoir is located northeast of the town of Limassol, Cyprus. The storage capacity of the reservoir is 13·6 × 10⁶ m³. The basin area of the Yermasoyia River, which feeds the reservoir, totals 122·5 km². This study aims to estimate the mean annual deposition amount in the reservoir, which originates from the corresponding basin. For the estimate of the mean annual sediment inflow into the reservoir, two mathematical models are used alternatively. Each model consists of three submodels: a rainfall‐runoff submodel, a soil erosion submodel and a sediment transport submodel for streams. In the first model, the potential evapotranspiration is estimated for the rainfall‐runoff submodel, and the soil erosion submodel of Schmidt and the sediment transport submodel of Yang are used. In the second model, the actual evapotranspiration is estimated for the rainfall‐runoff submodel, and the soil erosion submodel of Poesen and the sediment transport submodel of Van Rijn are used. The deposition amount in the reservoir is estimated by means of the diagram of Brune, which delivers the trap efficiency of the reservoir. Daily rainfall data from three rainfall stations, and daily values of air temperature, relative air humidity and sunlight hours from a meteorological station for four years (1986–89) were available. The computed annual runoff volumes and mean annual soil erosion rate are compared with the respective measurement data. Copyright © 2006 John Wiley & Sons, Ltd.     

MODELLING RAINFALL-RUNOFF FROM LARGE CATCHMENT TO BASIN SCALE: THE GOULBURN VALLEY,VICTORIA

A lumped parameter dynamic rainfall-runoff model, IHACRES, is applied to the large upland area (more than 4500 km²) of the Goulburn Valley Basin, Victoria, Australia to predict streamflow under different climatic conditions. This paper presents the first evaluation of a rainfall–runoff model at large catchment scale, which is comprehensive in terms of the number of catchments investigated and the number of calibration and simulation periods used. The basin is subdivided into 12 catchments (from 100 to 700 km²), each of which is calibrated separately. High values of model efficiency and low bias are consistently obtained for different calibration sub-periods for all catchments in the basin. Simulation or so-called validation tests are used to select the best models for each catchment. This allows simulation of the water regime during long historical (approximately 90 year) periods when only climatological (rainfall and temperature) data were available. This procedure is extremely important for the estimation of the effect of climate variability and of the possible impact of climate change on the hydrological regime in the region and, in particular, for supporting irrigation management of the basin. Analysis of a composite catchment (2417 km²) and its five separate subcatchments indicates that the information content in the rainfall–streamflow data is independent of catchment size. Dynamic modelling of the daily water balance at the macroscale is limited principally by the adequacy of the precipitation gauging network. When a good estimate of areal precipitation is available for a catchment, it is not necessary to consider subcatchment-scale variability for modelling if the only interest is the daily discharge and evaporation losses from the catchment.     

Effects of changes in canopy interception on stream runoff response and recovery following clear-cutting of a Japanese coniferous forest in Fukuroyamasawa Experimental Watershed in Japan

Understanding changes in evapotranspiration during forest regrowth is essential to predict changes of stream runoff and recovery after forest cutting. Canopy interception (Ic) is an important component of evapotranspiration, however Ic changes and the impact on stream runoff during regrowth after cutting remains unclear due to limited observations. The objective of this study was to examine the effects of Ic changes on long-term stream runoff in a regrowth Japanese cedar and Japanese cypress forest following clear-cutting. This study was conducted in two 1-ha paired headwater catchments at Fukuroyamasawa Experimental Watershed in Japan. The catchments were 100% covered by Japanese coniferous plantation forest, one of which was 100% clear-cut in 1999 when the forest was 70 years old. In the treated catchment, annual runoff increased by 301 mm/year (14% of precipitation) the year following clear-cutting, and remained 185 mm/year (7.9% of precipitation) higher in the young regrowth forest for 12–14 years compared to the estimated runoff assuming no clear-cutting. The Ic change was −358 mm/year (17% of precipitation) after cutting and was −168 mm/year (6.7% of precipitation) in the 12–14 years old regrowth forest compared to the observed Ic during the pre-cutting period. Stream runoff increased in all seasons, and the Ic change was the main fraction of evapotranspiration change in all seasons throughout the observation period. These results suggest that the change in Ic accounted for most of the runoff response following forest cutting and the subsequent runoff recovery in this coniferous forest.     

Hydrological trends and the evolution of catchment research in the Alptal valley,central Switzerland

When the observation of small headwater catchments in the pre-Alpine Alptal valley (central Switzerland) started in the late 1960s, the researchers were mainly interested in questions related to floods and forest management. Investigations of geomorphological processes in the steep torrent channels followed in the 1980s, along with detailed observations of biogeochemical and ecohydrological processes in individual forest stands. More recently, research in the Alptal has addressed the impacts of climate change on water supply and runoff generation. In this article, we describe, for the first time, the evolution of catchment research at Alptal, and present new analyses of long-term trends and short-term hydrologic behaviour. Hydrometeorological time series from the past 50 years show substantial interannual variability, but only minimal long-term trends, except for the ~2°C increase in mean annual air temperature over the 50-year period, and a corresponding shift towards earlier snowmelt. Similar to previous studies in larger Alpine catchments, the decadal variations in mean annual runoff in Alptal's small research catchments reflect the long-term variability in annual precipitation. In the Alptal valley, the most evident hydrological trends were observed in late spring and are related to the substantial change in the duration of the snow cover. Streamflow and water quality are highly variable within and between hydrological events, suggesting rapid shifts in flow pathways and mixing, as well as changing connectivity of runoff-generating areas. This overview illustrates how catchment research in the Alptal has evolved in response to changing societal concerns and emerging scientific questions.     

Spatial and temporal variability of the hydrological response in a small Mediterranean research catchment (Vallcebre,Eastern Pyrenees)

This paper analyses the spatial and temporal variability of the hydrological response in a small Mediterranean catchment (Cal Rodó). The first part of the analysis focuses on the rainfall–runoff relationship at seasonal and monthly scale, using an 8‐year data set. Then, using storm‐flow volume and coefficient, the temporal variability of the rainfall–runoff relationship and its relationship with several hydrological variables are analysed at the event scale from hydrographs observed over a 3‐year period. Finally, the spatial non‐linearity of the hydrological response is examined by comparing the Cal Rodó hydrological response with the Can Vila sub‐catchment response at the event scale. Results show that, on a seasonal and monthly scale, there is no simple relationship between rainfall and runoff depths, and that evapotranspiration is a factor that introduced some non‐linearity in the rainfall–runoff relationship. The analysis of monthly values also reveals the existence of a threshold in the relationship between rainfall and runoff depths, denoting a more contrasted hydrological response than the one usually observed in humid catchments. At the event scale, the storm‐flow coefficient has a clear seasonal pattern with an alternance between a wet period, when the catchment is hydrologically responsive, and a dry summer period, when the catchment is much less reactive to any rainfall. The relationship between the storm‐flow coefficient and rainfall depth, rainfall maximum intensity and base‐flow shows that observed correlations are the same as those observed for humid conditions, even if correlation coefficients are notably lower. Comparison with the Can Vila sub‐catchment highlights the spatial heterogeneity of the rainfall‐runoff relationship at the small catchment scale. Although interpretation in terms of runoff processes remains delicate, heterogeneities between the two catchments seem to be related to changes in the ratio between infiltration excess and saturation processes in runoff formation. Copyright © 2007 John Wiley & Sons, Ltd.