What would current ecosystems be like without the impact of mankind? This question, which is critical for ecosystem management, has long remained unanswered due to a lack of present-day data from truly undisturbed ecosystems. Using mountaineering techniques, we accessed pristine relict ecosystems in the Peruvian Andes to provide this baseline data and compared it with the surrounding accessible and disturbed landscape. We show that natural ecosystems and human impact in the high Andes are radically different from preconceived ideas. Vegetation of these ‘lost worlds’ was dominated by plant species previously unknown to science that have become extinct in nearby human-affected ecosystems. Furthermore, natural vegetation had greater plant biomass with potentially as much as ten times more forest, but lower plant diversity. Contrary to our expectations, soils showed relatively little degradation when compared within a vegetation type, but differed mainly between forest and grassland ecosystems. At the landscape level, a presumed large-scale forest reduction resulted in a nowadays more acidic soilscape with higher carbon storage, partly ameliorating carbon loss through deforestation. Human impact in the high Andes, thus, had mixed effects on biodiversity, while soils and carbon stocks would have been mainly indirectly affected through a suggested large-scale vegetation change.

Grasslands are very important regionally and globally because they store large amounts of carbon (C) and nitrogen (N) and provide food for grazing animals. Intensive degradation of alpine grasslands in recent decades has mainly impacted the upper root-mat/soil horizon, with severe consequences for nutrient uptake in these nutrient-limited ecosystems. We used ¹⁵N labeling to identify the role of individual soil layers for N-uptake by Kobresia pygmaea—the dominating plant in the degraded Tibetan pasture ecosystems. We hypothesized a very efficient N-uptake corresponding mainly to the vertical distribution of living roots (topsoil > subsoil). We assume that K. pygmaea develops a very dense root-mat, which has to be maintained by small aboveground biomass, to enable this efficient N-uptake. Consequently, a higher N-investment into roots compared to shoots was hypothesized. The ¹⁵N recovery in whole plants (~70%) indicated very efficient N-uptake from the upper injection depths (0–5 cm). The highest ¹⁵N amounts were recovered in root biomass, whereby ¹⁵N recovery in roots strongly decreased with depth. In contrast, ¹⁵N recovery in shoots was generally low (~18%) and independent of the ¹⁵N injection depth. This clearly shows that the low N demand of Kobresia shoots can be easily covered by N-uptake from any depth. Less living root biomass in lower versus upper soil was compensated by a higher specific activity of roots for N-uptake. The ¹⁵N allocation into roots was on average 1.7 times higher than that into shoots, which agreed well with the very high R/S ratio. Increasing root biomass is an efficient strategy of K. pygmaea to compete for belowground resources at depths and periods with available resources. This implies high C-costs to maintain root biomass (~6.0 kg DM m⁻²), which must be covered by a very low amount of photosynthetically active shoots (0.3 kg DM m⁻²). It also suggests that Kobresia grasslands react extremely sensitively toward changes in climate and management that disrupt this above-/belowground trade-off mechanism.