Supplementary material to "A conservative resource use strategy in agricultural grasslands counteracts lower productivity and water use efficiency under drought conditions"

Soil characterization MethodsSoil samples from a depth of 0-0.15 m were collected at three randomly selected locations per lysimeter, and these were then pooled to form a mix sample (approximately 500 g).These mix samples were air-dried, and sieved through a 2-mm sieve.Soil pH was measured potentiometrically in a solution of 0.01 M CaCl2 (Thomas, 1996).Soil organic matter (SOM) was determined by weight loss after ignition in a muffle furnace at 550°C for 7 hours (Schlichting and Blume, 1995).The soil texture was measured with an automated particle size analyser (PARIO; Meter Group; Munich, Germany).Before the analysis, 30 g of air-dried soil from each lysimeter were treated with 30 % H2O2 to remove organic matter and were chemically (0.15 M Na4P2O7) and mechanically (shaker) dispersed.The resulting suspension was analysed by PARIO, which calculates the particle size distribution by Stoke's law with a range spanning from 0.002 to 0.063 mm and provides a particle size distribution curve.Then, the suspension was flushed onto a cascaded sieve to obtain the particles between 0.063 and 2 mm.Particles smaller than 0.002 mm were calculated from the residual (i.e., unrecovered) mass.Particles were divided according to their size into sand ('large', 0.063 -2 mm), silt ('medium', 0.002 -0.063 mm) and clay ('small', < 0.002 mm) (after DIN 4022). ResultsRegarding soil parameters, SOM significantly decreased in all lysimeters by the end of the experiment from 6.03 ± 0.07 % to 4.88 ± 0.07 % (ANOVA, F = 240.408,p < 0.001), but no differences were seen between grassland types or irrigation levels.Soil pH (7.24 ± 0.03) did not change during the experiment.The analysis of soil particle size distribution revealed a sandy loam texture for all lysimeters.The particle size distribution in soil differed slightly between