The Isabena catchment (445 km2), Spain, features highly diverse spatial heterogeneity in land use, lithology and rainfall. Consequently, the relative contribution in terms of water and sediment yield varies immensely between its subcatchments, and also temporally. This study presents the synthesis of ~2.5 years of monitoring rainfall, discharge and suspended sediment concentration (SSC) in the five main subcatchments of the Isabena and its outlet. Continuous discharge at the subcatchment outlets, nine tipping bucket rainfall and automatic SSC samplers (complemented by manual samples), were collected from June 2011 until November 2013. The water stage records were converted to discharge using a rating curve derived with Bayesian regression. For reconstructing sediment yields, the data from the intermittent SSC sampling needed to be interpolated. We employed non-parametric multivariate regression (Quantile Regression Forests, QRF) using the discharge and rainfall data plus different aggregation levels of these as ancillary predictors. The subsequent Monte Carlo simulations allowed the determination of monthly sediment yields and their uncertainty. The stage–discharge rating curves showed wide credibility intervals for the higher stages, with great uncertainties associated with the discharge rates, especially during floods. The water yield of the subcatchments differed considerably. The entire catchment’s output was dominated by the northernmost subcatchment (~360 mm year−1). The smaller, southern subcatchments featured much higher variability and lower runoff rates (55–250 mm year−1). The SSCs exhibited a wide range and can exceed 100 g l−1 for the central subcatchments, where most of the badlands are located. For the reconstruction of the sedigraphs, the QRF method proved suitable with Nash–Sutcliffe indices of 0.50 to 0.84. The specific sediment yield ranges from relatively low (32 t km−2 year−1) in the highly vegetated north to high values (3,651 t km−2 year−1) in areas with many badland formations. The Isabena catchment shows high erosion dynamics with great variability in space and time, with stark contrasts even between adjacent subcatchments. The natural conditions make water and sediment monitoring and instrumentation very challenging; the measurement of discharge is particularly prone to considerable uncertainties. The QRF method employed for reconstructing sedigraphs and monthly yields proved well suited for the task.