Agroforestry parklands are the most widespread production systems across Burkina Faso and the Sahel, with key species being Vitellaria paradoxa and Parkia biglobosa. The overall aim of the research was to contribute to a better understanding of the interactions in the tree-crop associations, and the impact of Vitellaria paradoxa parklands on the water balance, in the current context of climate change. Specific goals were to (i) assess, through field experiments, the factors that limit the production of associated crops (Sorghum bicolor) in parkland systems, (ii) measure the transpiration of V. paradoxa on an hourly, seasonal and annual basis, and the variables predicting such transpiration, and (iii) quantify interception, runoff and infiltration at given distances from the V. paradoxa trunk both at the individual tree and parkland levels. The study of the tree-crop interaction has shown that under tree crowns complementary irrigation and the use of nitrogen and/or phosphorous have led to only a minor gain in the dry matter and sorghum grain yield. However, crown pruning has had a very significant effect on sorghum growth and production (with yields increasing by 348% to 520%), thereby suggesting that light is the limiting factor in the tree-sorghum association. Moreover, sorghum production under V. paradoxa has been significantly higher (by 56%) than under P. biglobosa. Measurements of transpiration vary considerably according to the time of the day, the months, the seasons and the years. Breast height diameter was a good predictor variable of sapwood area (R2=0.82), and accounts for 69% of such variation. Average transpiration per tree is established at 151 L day-1. In addition, the sap flux in the dry season (0.73 L h-1 dm-2) is significantly higher than in the rainy season (0.53 L h-1 dm-2). High hourly values of this flow are found in the daytime, reaching a peak at 5 PM (1.11 L h-1 dm-2). Except of rainfall and soil moisture, all climate variables are strongly correlated to the sap flow. Best regressions showed that active photosynthetic radiations and relative air humidity were the factors that best explained the sap flow (R2=0.46). The interception of rainfall by the V. paradoxa canopy varied according to the compass point and distance from the tree trunk. It was most significant in the West point (41.6%) and decreased away from the trunk. The greatest interception was registered near the trunk (32.5%) and the smallest one close to the crown (20.95%). On average 22.98% of rainfall was intercepted by V. paradoxa, and a notable logarithmic regression was established between the percentage of intercepted water and the amount of incidental rainfall. No major difference was found in the cumulative runoff in terms of distance from the V. paradoxa trunk, although runoff outside the crown was lower. Generally, cumulative runoff varied between 13 and 32% of incidental rainfall both under and outside the crown, with an average of 21.61%. Infiltrations at 4 and 10 m from the trunk were significantly higher than those measured at 1 and 40 m from the trunk. The lowest infiltration was measured at 1 m and the highest at 4 m from the trunk. Under the crown, infiltration increased away from the trunk before decreasing beyond the drip line. In general, cumulative infiltration varied proportionally to the rainfall level and accounted on average for 21.66% of incidental rainfall. Preserving a minimum of V. paradoxa trees in parklands is vital to ensure water infiltration and support agricultural production

Agroforestry parkland, Infiltration, Interactions trees/crops, Interception, Runoff, Limiting factor, Rainfall, Sap flow, P. biglobosa, V. paradoxa, water balance