In this paper we present the design, construction and performance of a self-contained float-controlled water level gauge for monitoring water levels in streams and small rivers. This device is inexpensive (cost of about EUR 220), easy to build (no electronics skills or specialized tools required; assembled in a few hours) and straightforward to use. The gauge remains autonomously operational for several weeks in remote locations without the need for an external power supply or solar panel and in the harsh tropical climatic conditions. Data can be downloaded wirelessly in the field using an Android smartphone or tablet. This gauge is capable of a measurement precision of ±1 mm at temperatures ranging from 20 to 30 °C and accurate to within 2 mm of manual readings in a controlled laboratory environment. In the field, the mean absolute error (MAE) of measurements taken with the water level gauge compared to that obtained with the OTT-SE200 - a commercial float-controlled angle encoder water level gauge - over a full tropical rainy season and for a measurement range of 0.5 m, was 2.6 mm (n = 8,017).

Riparian grass buffers reduce the velocity of water flowing over the soil surface during storms, capturing surface runoff (SR) and trapping soil particles eroded from cultivated slopes. Rarely quantified under steep slope conditions (>45 %), this phenomenon probably occurs in many mountain agroecosystems in the humid tropics. In Southeast Asia, teak plantations are often established on steep slopes where they can lead to significant soil loss (SL), particularly when the understory is removed. Therefore, we aimed to: quantify the effect of riparian grass buffers on SR and SL downstream of a teak plantation; and estimate the trapping efficiency of riparian grass buffers for water (WTE) and sediment (STE). Field measurements were carried out in Northern Lao PDR during the2014 rainy season, considering riparian zones with contrasted ground covers: (1) uncovered (URZ - 7-year-old teak trees with mean grass and litter densities of 7 % (SD 2 %) and 4 % (SD 3 %), respectively; (2) transitional (TRZ - 7-year-old teak trees with mean grass and litter densities of 19 % (SD 10 %) and 56 % (SD 9 %), respectively; and (3) covered (CRZ - grassed areas without teak trees with mean grass and litter densities of 46 % (SD 13 %) and 47 % (SD 21 %), respectively). WTE and STE were estimated based on measurements carried out from 6 July to 6 September2014 under natural rainfall conditions, using pairs of triplicate Gerlach troughs installed at the upper and lower margins of 5- and 10-m riparian sections (encompassing areas of similar to 25 and 50 m(2), respectively). Runoff coefficient (Rc), SL, and soil surface features were measured on the occasion of 20 rainfall events in 1-m(2) microplots. Rc and SL were higher in URZ (56 %, 5791 gm(-2)) than in TRZ (13 %, 250 gm(-2)) and CRZ (19 %, 159 gm(-2)). Median WTE and STE were the highest along the 10-m TRZ + CRZ riparian grass buffer at 85 % and 97 % respectively. Partial Least Square Regression (PLSR) modelling yielded a good agreement between observation and prediction for WTE and STE at seasonal scale. Overall, the results of this work indicate that 5 to 10 m riparian grass buffers limit the export of surface water and sediment downstream during small (24-h rainfall <= 20.9 mmd(-1), return period <= 1 year) to large storms (40.0 mmd(-1) < 24-h rainfall <= 84.5 mmd(-1), 1.01 year

It is well known that biopores are crucial for soil functioning. However, their dynamics is rarely studied and their origin with regards to the soil organisms involved is still hard to determine. In this study we investigated the diversity of biopores and their regeneration rates in situ in various pedoclimatic conditions. Our approach involved field incubation of repacked soil cores with lateral openings across nine study sites in five countries (France, Vietnam, India, Laos and Thailand). After 12 months, biopores were characterized by X-ray computed tomography and grouped according to their diameter, length and sphericity index using principal component analysis followed by K-means clustering. The regeneration dynamics of biopores was assessed by comparing those created after one year of incubation to the biopores determined in soil cores taken from the surrounding soils (assuming the latter are in a steady-state). Additionally, we examined the relationships between newly formed biopores and soil macrofauna taxa. Our results evidenced significant variability in biopore diameter (0.90 to 15.84 mm), length (1 to 1600 mm) and sphericity index (0.03 to 0.93). We propose 10 biopore groups allowing to distinguish most of the study sites. Complete regeneration of biopores after 12 months was achieved in seven out of nine sites. Three groups of biopores showed a positive relation with earthworm abundance (r values ranged from 0.69 to 0.90), whereas the other groups of biopores showed no association with any macrofauna taxa. We conclude that biopore formation can be assessed under field conditions with repacked soil cores, regardless the pedoclimatic conditions. However, the involvement of macrofauna other than earthworms in biopore formation still remains to be unraveled. To capture their contribution to biopore formation, improvements of the repacked soil core approach and complementary laboratory experiments were suggested.

Fecal bacteria in surface water may indicate threats to human health. Our hypothesis is that village settlements in tropical rural areas are major hotspots of fecal contamination because of the number of domestic animals usually roaming in the alleys and the lack of fecal matter treatment before entering the river network. By jointly monitoring the dynamics of Escherichia coli and of seven stanol compounds during four flood events (July-August 2016) at the outlet of a ditch draining sewage and surface runoff out of a village of Northern Lao PDR, our objectives were (1) to assess the range of E. coli concentration in the surface runoff washing off from a village settlement and (2) to identify the major contributory sources of fecal contamination using stanol compounds during flood events. E. coli pulses ranged from 4.7 × 104 to 3.2 × 106 most probable number (MPN) 100 mL-1, with particle-attached E. coli ranging from 83 to 100%. Major contributory feces sources were chickens and humans (about 66 and 29%, respectively), with the highest percentage switching from the human pole to the chicken pole during flood events. Concentrations indicate a severe fecal contamination of surface water during flood events and suggest that villages may be considered as major hotspots of fecal contamination pulses into the river network and thus as point sources in hydrological models.

Despite the large contribution of macropores made by soil engineers to the soil macroporosity and water infiltration, few studies have addressed the specific contribution of soil engineer groups, dynamics of biopores and their efficiency in conducting water. Thus, we aimed to investigate the link between soil macrofauna, soil biopores and water infiltration under different pedoclimatic conditions. To do so, we conducted an experimentation in twelve study sites with a large longitudinal gradient from France to Vietnam. The experiment consisted in the field incubation of repacked soil in cores (15 cm in height and 15 cm in diameter) and controlling the activity of soil engineers in the manner of litter bag. For each site, soil columns were: (i) covered with a mesh (200µm) or not and (ii) with or without addition of organic residues to the soil surface. After 12 months, we measured (i) the 3D organization of biopores by X-ray computed tomography and (ii) the saturated hydraulic conductivity by Beerkan method. In addition, soil macrofauna communities and the 3D organization of biopores was measured in each study field. Addition of organic residues increased up to 2-fold the volume percentage of biopores which reached similar values than those observed for each study field. The co-inertia analysis between the data matrix characterizing the shape of biopores and the data matrix of the macrofauna communities showed no statistically significant correlation. Saturated hydraulic conductivity increased with the presence of biopores by 2 to 50-fold with the lowest increased in soils presenting largest saturated hydraulic conductivity. In conclusion, these results demonstrated that biopores are rapidly regenerated regardless the pedoclimatic conditions while the efficiency of biopores in conducting water is related to soil properties.

The environmental distribution of Burkholderia pseudomallei, the causative agent of melioidosis, remains poorly understood. B. pseudomallei is known to have the ability to occupy a variety of environmental niches, particularly in soil. This paper provides novel information about a putative association of soil biogeochemical heterogeneity and the vertical distribution of B. pseudomallei. We investigated (1) the distribution of B. pseudomallei along a 300-cm deep soil profile together with the variation of a range of soil physico-chemical properties; (2) whether correlations between the distribution of B. pseudomallei and soil physico-chemical properties exist and (3) when they exist, what such correlations indicate with regards to the environmental conditions conducive to the occurrence of B. pseudomallei in soils. Unexpectedly, the highest concentrations of B. pseudomallei were observed between 100 and 200 cm below the soil surface. Our results indicate that unravelling the environmental conditions favorable to B. pseudomallei entails considering many aspects of the actual complexity of soil. Important recommendations regarding environmental sampling for B. pseudomallei can be drawn from this work, in particular that collecting samples down to the water table is of foremost importance, as groundwater persistence appears to be a controlling factor of the occurrence of B. pseudomallei in soil.

As flood events are expected to become more frequent due to climate change, investigating how overland flow exports terrestrial nutrients, carbon and living organisms into aquatic systems is essential for understanding both soil and stream ecosystem status. Here we assessed how dissolved organic carbon (DOC), total suspended sediments (TSS), and stream bacterial diversity responded to stream discharge and overland flow during stormflow in a tropical catchment. A higher humification index and a decreasing ratio of allochthonous to autochthonous DOC indicated that DOC from soils was exported to stream during the flood. The δ13C and δ15N of particulate matter was indicative of a source in the cultivated areas of the upper catchment and of subsurface soils (stream banks and gullies) in the downstream section. Bacterial richness of particle-attached (PA) and the free-living (FL) fractions increased with the flood progression in the upstream section. Moreover, the community structure of the PA fraction in the stream was more similar to that of overland flow than was the FL fraction. This suggests that the soil PA bacterial community was washed-out with overland flow during the flood recession. The relative contribution of sources and the composition of TSS, rather than hydrological regime, significantly drove the composition of bacterial community. In conclusion, our results emphasize that overland flow during a flood event strongly influences the structure of stream bacterial communities further underlining the biological connectivity between terrestrial runoff and stream flow.

Bacterial pathogens in surface waters may threaten human health, especially in developing countries, where untreated surface water is often used for domestic needs. The objective of the long-term multiscale monitoring of Escherichia coli ([E. coli]) concentration in stream water, and that of associated variables (temperature (T), electrical conductance (EC), dissolved oxygen concentration ([DO]) and saturation (DO%), pH (pH), oxidation-reduction potential (ORP), turbidity (Turb), and total suspended sediment concentration ([TSS])), was to identify the drivers of bacterial dissemination across tropical catchments. This data description paper presents three datasets (see “Data availability” section) collected at 31 sampling stations located within the Mekong River and its tributaries in Lao PDR (0.6–25 946 km2) from 2011 to 2021. The 1602 records have been used to describe the hydrological processes driving in-stream E. coli concentration during flood events, to understand the land-use impact on bacterial dissemination on small and large catchment scales, to relate stream water quality and diarrhea outbreaks, and to build numerical models. The database may be further used, e.g., to interpret new variables measured in the monitored catchments, or to map the health risk posed by fecal pathogens.

Abstract. Contamination of surface waters through microbiological pollutants is a major concern to public health. Although long-term and high-frequency E. coli monitoring can help prevent diseases from fecal pathogenic microorganisms, this monitoring is time consuming and expensive. Process-driven models are an alternative method for determining fecal pathogenic microorganisms. However, process-based modeling still has limitations in improving the model accuracy because of the complex mechanistic relationships among hydrological and environmental variables. On the other hand, with the rise in data availability and computation power, the use of data-driven models is increasing. Therefore, in this study, we simulated the transport of Escherichia coli (E. coli) in a 0.6 km² tropical headwater catchment located in Lao PDR using a deep learning model and a process-based model. The deep learning model was built using the long short-term memory (LSTM) technique, whereas the process-based model was constructed using the Hydrological Simulation Program–FORTRAN (HSPF). First, we calibrated both models for surface as well as for subsurface flow. Then, we simulated the E. coli transport with 6 min time steps with both the HSPF and LSTM models. The LSTM provided accurate results for surface and subsurface flow, by showing 0.51 and 0.64 of Nash–Sutcliffe Efficiency (NSE), respectively, whereas the NSE values yielded by the HSPF were −0.7 and 0.59 for surface and subsurface flow. The simulated E. coli concentration from LSTM also improved, yielding an NSE of 0.35, whereas the HSPF showed an unacceptable performance, with an NSE value of −3.01. This is because of the limitation of HSPF in capturing the dynamics of E. coli with land-use change. The simulated E. coli concentration showed rise and drop patterns corresponding to annual changes in land use. This study shows the application of deep learning-based models as an efficient alternative to process-based models for E. coil fate and transport simulation at the catchment scale.

Surface water contamination by pathogen bacteria remains a threat to public health in the rural areas of developing countries. Fecal indicator bacteria (FIB) like Escherichia coli (E. coli) are widely used to assess water contamination, but their behavior in tropical ecosystems is poorly documented. Our study focused on headwater wetlands which are likely to play a key role in stream water purification of fecal pollutants. Our main objectives were to: (i) evaluate decay rates (k) of the total, particle-attached and free-living E. coli; (ii) quantify the relative importance of solar radiation exposition and suspended particles deposition on k; and (iii) investigate E. coli survival in the deposited sediment. We installed and monitored 12 mesocosms, 4500 mL each, across the main headwater wetland of the Houay Pano catchment, northern Lao People’s Democratic Republic (Lao PDR), for 8 days. The four treatments with triplicates were: sediment deposition-light (DL); sediment deposition-dark (DD); sediment resuspension-light (RL); and sediment resuspension-dark (RD). Particle-attached bacteria predominated in all mesocosms (97 ± 6%). Decay rates ranged from 1.43 ± 0.15 to 1.17 ± 0.13 day−1 for DL and DD treatments, and from 0.50 ± 0.15 to −0.14 ± 0.37 day−1 for RL and RD treatments. Deposition processes accounted for an average of 92% of E. coli stock reduction, while solar radiation accounted for around 2% over the experiment duration. The sampling of E. coli by temporary resuspension of the deposited sediment showed k values close to zero, suggesting potential survival or even growth of bacteria in the sediment. The present findings may help parameterizing hydrological and water quality models in a tropical context.

Bacterial pathogens in surface waters threaten human health. The health risk is especially high in developing countries where sanitation systems are often lacking or deficient. Considering twelve flash-flood events sampled from 2011 to 2015 at the outlet of a 60-ha tropical montane headwater catchment in Northern Lao PDR, and using Escherichia coli as a fecal indicator bacteria, our objective was to quantify the contributions of both surface runoff and sub-surface flow to the in-stream concentration of E. coli during flood events, by (1) investigating E. coli dynamics during flood events and among flood events and (2) designing and comparing simple statistical and mixing models to predict E. coli concentration in stream flow during flood events. We found that in-stream E. coli concentration is high regardless of the contributions of both surface runoff and sub-surface flow to the flood event. However, we measured the highest concentration of E. coli during the flood events that are predominantly driven by surface runoff. This indicates that surface runoff, and causatively soil surface erosion, are the primary drivers of in-stream E. coli contamination. This was further confirmed by the step-wise regression applied to instantaneous E. coli concentration measured in individual water samples collected during the flood events, and by the three models applied to each flood event (linear model, partial least square model, and mixing model). The three models showed that the percentage of surface runoff in stream flow was the best predictor of the flood event mean E. coli concentration. The mixing model yielded a Nash-Sutcliffe efficiency of 0.65 and showed that on average, 89% of the in-stream concentration of E. coli resulted from surface runoff, while the overall contribution of surface runoff to the stream flow was 41%. We also showed that stream flow turbidity and E. coli concentration were positively correlated, but that turbidity was not a strong predictor of E. coli concentration during flood events. These findings will help building adequate catchment-scale models to predict E. coli fate and transport, and mapping the related risk of fecal contamination in a global changing context.

Mountain regions of the humid tropics are characterized by steep slopes and heavy rains. These regions are thus prone to both high surface runoff and soil erosion. In Southeast Asia, uplands are also subject to rapid land‐use change, predominantly as a result of increased population pressure and market forces. Since 1998, the Houay Pano site, located in northern Lao PDR (19.85°N 102.17°E) within the Mekong basin, aims at assessing the long‐term impact of the conversion of traditional slash‐and‐burn cultivation systems to commercial perennial monocultures such as teak tree plantations, on the catchment hydrological response and sediment yield. The instrumented site monitors hydro‐meteorological and soil loss parameters at both microplot (1 m 2 ) and small catchment (0.6 km 2 ) scales. The monitored catchment is part of the network of critical zone observatories named Multiscale TROPIcal CatchmentS (M‐TROPICS). The data shared by M‐TROPICS in Houay Pano are (1) rainfall, (2) air temperature, air relative humidity, wind speed, and global radiation, (3) catchment land use, (4) stream water level, suspended particulate matter, bed particulate matter and stones, (5) soil surface features, and (6) soil surface runoff and soil detachment. The dataset has already been used to interpret suspended particulate matter and bed particulate matter sources and dynamics, to assess the impact of land‐use change on catchment hydrology, soil erosion, and sediment yields, to understand bacteria fate and weed seed transport across the catchment, and to build catchment‐scale models focused on hydrology and water quality issues. The dataset may be further used to, for example, assess the role of headwater catchments in large tropical river basin hydrology, support the interpretation of new variables measured in the catchment (e.g., contaminants other than faecal bacteria), and assess the relative impacts of both climate and land‐use change on the catchment.

Organic amendments are believed to help increase the soil carbon storage and therefore improve soil quality, which may be important in the context of climate change. However, the added value of organic amendments for farmers must be clearly demonstrated in order to convince them of the utility of their use. The aims of this study were: (i) to investigate the impact on maize of compost and vermicompost combined with two levels (negligible and significant) of plant water stress; and (ii) to determine how the organic amendments affected the soil's physical properties and maize productivity. Water stress levels were imposed by controlling the matric potential of soil columns in which cultivated soil characteristics was mimicked (10 cm topsoil with organic amendments, above a 50 cm subsoil without any inputs containing the majority of the roots). Plant and soil characteristics were monitored daily for 70 days. Our results show that the use of organic amendments is profitable for farmers as: (i) maize performances were increased in both moisture regimes; and (ii) the improvement was particularly striking in terms of yield. No additional benefits were measured when using vermicompost instead of compost. The data suggest that the improvement in plant characteristics did not result from increased water storage in the soils with organic amendments, but rather from better access to the water, resulting in faster root development in the macroporosity of the amended soils.

Many mountainous regions of the humid tropics experience serious soil erosion following rapid changes in land use. In northern Lao People’s Democratic Republic (PDR), the replacement of traditional crops by tree plantations, such as teak trees, has led to a dramatic increase in floods and soil loss and to the degradation of basic soil ecosystem services such as water filtration by soil, fertility maintenance, etc. In this study, we hypothesized that conserving understory under teak trees would protect soil, limit surface runoff, and help reduce soil erosion. Using 1 m2 microplots installed in four teak tree plantations in northern Lao PDR over the rainy season of 2017, this study aimed to: (1) assess the effects on surface runoff and soil loss of four understory management practices, namely teak with no understory (TNU; control treatment), teak with low density of understory (TLU), teak with high density of understory (THU), and teak with broom grass, Thysanolaena latifolia (TBG); (2) suggest soil erosion mitigation management practices; and (3) identify a field visual indicator allowing a rapid appraisal of soil erosion intensity. We monitored surface runoff and soil loss, and measured teak tree and understory characteristics (height and percentage of cover) and soil surface features. We estimated the relationships among these variables through statistics and regression analyses. THU and TBG had the smallest runoff coefficient (23% for both) and soil loss (465 and 381 g·m−2, respectively). The runoff coefficient and soil loss in TLU were 35% and 1115 g·m−2, respectively. TNU had the highest runoff coefficient and soil loss (60%, 5455 g·m−2) associated to the highest crusting rate (82%). Hence, the soil loss in TBG was 14-times less than in TNU and teak tree plantation owners could divide soil loss by 14 by keeping understory, such as broom grass, within teak tree plantations. Indeed, a high runoff coefficient and soil loss in TNU was explained by the kinetic energy of rain drops falling from the broad leaves of the tall teak trees down to bare soil, devoid of plant residues, thus leading to severe soil surface crusting and soil detachment. The areal percentage of pedestal features was a reliable indicator of soil erosion intensity. Overall, promoting understory, such as broom grass, in teak tree plantations would: (1) limit surface runoff and improve soil infiltrability, thus increase soil water stock available for both root absorption and groundwater recharge; and (2) mitigate soil loss while favoring soil fertility conservation.

Recent increase in climate change has resulted in rise of hydrologic extreme events, which demands better understanding of flow patterns in catchment. Modeling surface and sub-surface flow at high temporal resolution helps to understand catchment dynamics. In this study, we simulated surface and sub-surface flow in a Laotian catchment at 6-minute resolution. We used one physically based model called Hydrological Simulated Program-FORTRAN (HSPF) and developed two deep learning-based models. One deep learning model consisted of only one long short-term memory (LSTM), whereas the other model simulated processes in each hydrologic response unit (HRU) by defining one separate LSTM for each HRU. The models consider environmental data as well as changing landuse in catchment and predict surface and sub-surface flows. Our results show that simple LSTM model outperformed other models for surface runoff prediction, whereas the HRU-based LSTM model better predicted patterns and slopes in sub-surface flow in comparison with other models.

Humid tropical mountainous area experiences serious soil erosion due to rapid changes in landuse, sometimes implying erosion prone management practices. In this study, we hypothesized that keeping understorey in teak tree plantation would protect soil and avoid soil erosion. We assessed the effects of 4 management practices in teak tree plantation area on water and soil losses using 6 replicated 1-m2 microplots in four plantations situated in Northern Laos during the wet season of 2017. The landuses in the four plantations were teak without understorey (TNU), teak with low density of understorey (TLU), teak with high density of understorey (THU), and teak with broom grass, Thysanolaena latifolia (TBG). During the wet season of 2017, we monitored surface runoff and soil loss for 22 rainfall events. We also measured some of the teak tree and understorey characteristics (i.e. height and percentage of cover) and the percentage areas of soil surface features (i.e. litter, free aggregates, crusting, etc.). Relationships among these variables was estimated through multiple statistics and regression analyses. We found that runoff coefficient and soil loss were the smallest for THU and TBG: runoff coefficient was 23% for both treatments, and soil losses were 465 and 381 g m-2, respectively. Runoff coefficient and soil loss for TLU were 35% and 1115 g m-2, respectively. We observed the highest runoff coefficient and soil loss under TNU (60%, 5455 g m-2) associated to the highest crusting rate (82%). High runoff coefficient and soil loss under TNU was explained by the kinetic energy of rain drops falling from the broad leaves of the tall teak trees down to bare soil, devoid of plant residues, thus leading to severe soil surface crusting and detachment. Overall, promoting understorey such as broom grass in teak tree plantations would (1) limit surface runoff and improve soil infiltrability, thus increase the soil water stock available for both root absorption and groundwater recharge, and (2) mitigate soil loss and favour soil fertility conservation.

The French Critical Zone research infrastructure, OZCAR-RI, gathers 20 observatories sampling various compartments of the critical zone, each having developed their own data management and distribution systems. A common information system (Theia/OZCAR IS) was built to make theirin situobservation FAIR (findable, accessible, interoperable, reusable). The IS architecture was designed after consultation of the users, data producers and IT teams involved in data management. A common data model based on various metadata standards was defined to create information fluxes between observatories' ISs and the Theia/OZCAR IS. Controlled vocabularies were defined to develop a data discovery web portal offering a faceted search with various criteria, including variables names and categories that were harmonized in a thesaurus published on the web. This paper describes the IS architecture, the pivot data model and open-source solutions used to implement data discovery, and future steps to implement data downloading and interoperability services.

In Montane Southeast Asia, deforestation and unsuitable combinations of crops and agricultural practices degrade soils at an unprecedented rate. Typically, smallholder farmers gain income from “available” land by replacing fallow or secondary forest by perennial crops. We aimed to understand how these practices increase or reduce soil erosion. Ten land uses were monitored in Northern Laos during the 2015 monsoon, using local farmers' fields. Experiments included plots of the conventional system (food crops and fallow), and land uses corresponding to new market opportunities (e.g. commercial tree plantations). Land uses were characterized by measuring plant cover and plant mean height per vegetation layer. Recorded meteorological variables included rainfall intensity, throughfall amount, throughfall kinetic energy (TKE), and raindrop size. Runoff coefficient, soil loss, and the percentage areas of soil surface types (free aggregates and gravel; crusts; macro-faunal, vegetal and pedestal features; plant litter) were derived from observations and measurements in 1-m2 micro-plots. Relationships between these variables were explored with multiple regression analyses. Our results indicate that TKE induces soil crusting and soil loss. By reducing rainfall infiltration, crusted area enhances runoff, which removes and transports soil particles detached by splash over non-crusted areas. TKE is lower under land uses reducing the velocity of raindrops and/or preventing an increase in their size. Optimal vegetation structures combine minimum height of the lowest layer (to reduce drop velocity at ground level) and maximum coverage (to intercept the largest amount of rainfall), as exemplified by broom grass (Thysanolaena latifolia). In contrast, high canopies with large leaves will increase TKE by enlarging raindrops, as exemplified by teak trees (Tectona grandis), unless a protective understorey exists under the trees. Policies that ban the burning of multi-layered vegetation structure under tree plantations should be enforced. Shade-tolerant shrubs and grasses with potential economic return could be promoted as understorey.

Groundwater contribution to streamflow sustains biodiversity and enhances ecosystem services, especially under monsoon-driven climate where stream baseflow is often the only available water resource during the dry season. We assessed how land use change influences streamflow and its groundwater contribution in a small headwater catchment subject to shifting cultivation in Montane Southeast Asia. Continuous time series of rainfall, reference evapotranspiration, groundwater level, stream discharge and electrical conductivity (EC) of surface and groundwater were monitored from 2002 to 2007. With the rainfall-runoff model GR4J, we investigated temporal changes in the hydrological behaviour of the study catchment to verify consistencies with observed land use change. An EC-based hydrograph separation method allowed estimating the groundwater contribution to 104 stormflow events. Mean soil surface crusting rates corresponding to each of the nine land uses identified in the catchment were determined using 236 standard 1-m(2) micro-plots. Mean plant cover for each land use was assessed in 10 x 10-m(2) plots. Bedrock topography and soil layers' structure were assessed by electrical resistivity tomography to determine pathways of subsurface storm flows. Our results indicate that an increase in the catchment's areal percentage of fallow from 33% to 71% led to a decrease in the annual runoff coefficient from 43% to 26%. The concurrent reduction of soil crusting rate over the catchment, from 48% to 30%, increased rainwater infiltration. Consecutively, groundwater contribution to storm streamflow increased from 83% to 94%, highlighting the protective role of a dense vegetation cover against flash floods. The overall reduction of the annual basin water yield for inter-storm streamflow from 450 to 185 nun suggests that the potential gain in groundwater recharge was offset by the increased root water uptake for evapotranspiration, as confirmed by the drop in the groundwater level. This analysis illustrates how two different land uses with opposite impacts on soil permeability (i/ extensive soil surface crusting under annual crops resulting in limited runoff infiltration or ii/ fallow regrowth promoting both infiltration and evapotranspiration) both inhibit groundwater recharge. The maintenance of strips of fallow buffers between annual crop plots can slow down runoff and locally promote infiltration and groundwater recharge while limiting evapotranspiration.