Vegetation dynamics and erosion resistance of sea dyke grassland

Abstract

The high erosion resistance of grass swards on c1ay dykes seems to be largely determined by the amount of roots and the root system architecture. Following on previous studies on river dykes which revealed high root densities and erosion resistance in unfertilised pastures and hay-meadows, we experimentally investigated the influence of cessation of fertiliser application in combination with different management regimes on factors affecting water erosion of sea dyke grassland in the NetherIands. From 1991 to 1994 a study was done on the botanical composition, vegetation cover, above- and belowground biomass production (root density) and chemical composition of the soil and plant tissue.The erodibility of the sward was tested in field and laboratory experiments, and shear strength was measured. Unfertilised, species-rich grasslands were investigated as a reference. After continuing the experiment till 1997, data on biomass production and chemical composition in 1997 were analysed as well. Three to four years after cessation of fertilisation we observed a slight increase in species number (from 13 to 16 species per 25 m'), a decline in cover or abundance of Lolium perenne and an increase of species of less nutrient-rich conditions. Also, a rather rapid dec1ine was detected in the aboveground biomass production from 9-10 tonnes dry matter to 7 t hay' after 4 years, and subsequently to 5-6 t hay' after seven years. Shifting from grazing to haymaking without the use of fertiliser resulted in the most drastic drop in biomass, reaching the level ofthe reference dykes. Together with the dec1ine in biomass, a drop in mineral nitrogen availability was measured . The total soil N pool, however, was not changed. No dec1ine in the cover of the vegetation (cut back to a height of 2 cm) or shoot density was found. In the mown treatments on both pastures and hayfields, cessation of fertilisation resulted in an increased root length and root weight, and lower shoot to root ratios. Species-rich old meadows exhibited relatively high root length and root weight values, and low shoot to root ratio's, and appeared to be characterised by a less steep decline of root density with depth than the species-poor fertilised pastures. Pastures were observed to have somewhat high specific root lengths (SRL, thin roots), and hayfields low SRL (thick roots), except for the species rich hay-meadow, characterized by high SRL. No seasonal fluctuation in root mass was measured in both fertilised pastures and unfertilised old hay meadows, so that the root length fluctuation measured in both types is caused by growth and decay of fine roots. Spatial variation in root weights seemed to be higher in the floristically homogeneous fertilised pastures. Shear strength measurements did not reflect soil cohesion due to higher root density and therefore the use of shear strength is not recommended as a parameter for erosion resistance. Although erosion sprayexperiments did not show significant effects between treatments, the erosion centrifuge experiments showed a significant relation between resistance against intemal erosion and the combined effect ofroot density and particIe size. In dyke grassland management we conclude that the main prerequisites for a high erosion resistance of the sod constitute avoiding the use of fertiliser in combination with haymaking or grazing and proper maintenance sustained over long periods. While judging the erodibility, more emphasis should be put on botanical composition than on clay quality.