International audienceAdequate water management is required to im- prove the efficiency and sustainability of agricultural sys- tems when water is scarce or over-abundant, especially in the case of land use changes. In order to quantify, to pre- dict and eventually to control water and solute transport into soil, soil hydraulic properties need to be determined pre- cisely. As their determination is often tedious, expensive and time-consuming, many alternative field and laboratory tech- niques are now available. The aim of this study was to de- termine unsaturated soil hydraulic properties under different land uses and to compare the results obtained with differ- ent measurement methods (Beerkan, disc infiltrometer, evap- oration, pedotransfer function). The study has been realized on a tropical sandy soil in a mini-watershed in northeastern Thailand. The experimental plots were positioned in a rub- ber tree plantation in different positions along a slope, in ruzi grass pasture and in an original forest site. Non-parametric statistics demonstrated that van Genuchten unsaturated soil parameters (Ks, α and n) were significantly different ac- cording to the measurement methods employed, whereas the land use was not a significant discriminating factor when all methods were considered together. However, within each method, parameters n and α were statistically different ac- cording to the sites. These parameters were used with Hy- drus1D for a 1-year simulation and computed pressure head did not show noticeable differences for the various sets of parameters, highlighting the fact that for modeling, any of these measurement methods could be employed. The choice of the measurement method would therefore be motivated by the simplicity, robustness and its low cost

SUMMARYIn order to get the benefit of the growing world demand for natural rubber, Hevea brasiliensis is increasingly planted in drought-prone areas, such as in the southern part of northeast Thailand. Modelling can be a useful approach in identifying key points of improvement for rubber tree cultivation in such water-limited areas. The first objective of this study was to test the possibility of using the Water Nutrients and Light Capture in Agroforestry Systems (WaNuLCAS) model as an exploratory tool to simulate water use, growth and latex production in a pure stand on a daily basis. The second was to evaluate the relative accuracy of predictions with the current model version. Finally, the third aim of this study was to identify particular parameterisations that may be adapted to improve overall prediction quality. Model outputs were compared to measurements recorded in a mature rubber tree stand of RRIM 600 clones growing in the water-limited area of northeast Thailand. The period of analysis concerned seven months of full foliation, from May to November, including a severe drought spell. Whole-tree transpiration was estimated by xylem sap flow measurement from 11 trees. The results show that the model was able to simulate daily and seasonal change of soil water content, tree transpiration, girth increment and latex production within plausible ranges. However, under detailed scrutiny, the predictions show large inaccuracies compared to the observations: soil water content (determination coefficient (R2) = 0.461, relative root mean square error (RMSErel) = 35%), tree transpiration (R2 = 0.104, RMSErel = 94%), tree girth increment (R2 = 0.916, RMSErel = 208%) and latex production (R2 = 0.423, RMSErel = 169%). As soil water content was overestimated during the driest periods, no water stress was predicted and transpiration, growth and latex production were logically overestimated during such periods. However, tree transpiration was also largely overestimated in conditions of non-limiting soil water availability with high evaporative demand. Hence, two key points of parameterisation and improvement are identified for better simulation in our conditions: the soil water balance and particularly the ratio between water infiltration and run-off, and the regulation of transpiration under high evaporative demand. In conclusion, the WaNuLCAS model is usable as an exploratory model to simulate water use, growth and production for a pure rubber tree stand. However, in our conditions of much degraded soil and high evaporative demand, the modules of soil water balance and tree transpiration require particular parameterisations and improvement.

Abstract. Adequate water management is required to improve the efficiency and sustainability of agricultural systems when water is scarce or over-abundant, especially in the case of land-use changes. In order to quantify, to predict and eventually to control water and solute transport into soil, soil hydraulic properties need to be determined precisely. As their determination is often tedious, expensive and time-consuming, many alternative field and laboratory techniques are now available. The aim of this study was to determine unsaturated soil hydraulic properties under different land-uses and to compare the results obtained with different measurement methods (Beerkan, Disk infiltrometer, Evaporation, pedotransfer function). The study has been realised on a tropical sandy soil in a mini watershed in NE Thailand. The experimental plots were positioned in a rubber tree plantation in different positions along a slope, in ruzi grass pasture and in an original forest site. Non parametric statistics demonstrated that van Genuchten unsaturated soil parameters (Ks, α and n), were significantly different according to the measurement methods employed whereas location was not a significant discriminating factor when all methods were considered together. However within each method, parameters n and α were statistically different according to the sites. These parameters were used with Hydrus1D for a one year simulation and computed pressure head did not show noticeable differences for the various sets of parameters, highlighting the fact that for modelling, any of these measurement method could be employed. The choice of the measurement method would therefore be motivated by the simplicity, robustness and its low cost.

Abstract. Infiltration into water repellent soils has been widely observed, quantified and documented. The modelling of water infiltration into water repellent soils is more rarely taken into account explicitly. In this study, we modelled water infiltration into water repellent soils considering explicitly the contact angle, with the geometrical pore model proposed and validated previously. The applied microscopical approach showed good agreement with macroscopical models and with experimental data. We firstly investigated the case of contact angles lower than 90°, for the cylindrical pore and pearl necklace (PN) models. The cumulative infiltrations were numerically generated versus contact angle and for different pore radii. Then, the modelled infiltration curves were fitted to the two-terms Philip equation and parameters S and A were evaluated versus contact angle. As predicted sorptivity S decreased with increasing contact angle, and the constant infiltration rate A increased with contact angle for both models. Then, the modelled data were fitted to the numerical solution of the Richards equation to derive the equivalent hydraulic parameters assuming van Genuchten model. The results showed that the contact angle decreased the saturated hydraulic conductivity and increased the parameter α. Lastly, our model was used to investigate strong water repellency with contact angles higher than 90°. Cumulative infiltration and related Philip parameters, S and A, were evaluated versus water pressure head at surface h0 and contact angles (between 90° and 96°). Our model may be used to predict water infiltration into water repellent soils for both moderate and strong water repellency, including fingering features

No abstract