Can We Detect Changes in Amazon Forest Structure Using Measurements of the Isotopic Composition of Precipitation?

• Published in: Geophysical research letters 2019-12, Vol.46 (24), p.14807-14816
• Main Authors: Pattnayak, K.C., Tindall, J .C., Brienen, R. J. W., Barichivich, J., Gloor, E.
• Format: Article
• Language: English
• Subjects: Air parcel trajectories; Air parcels; Atmosphere; Change detection; Climate models; Computer simulation; Deforestation; Dry season; Evapotranspiration; Forests; Gradients; Heavy water; Hydrologic cycle; Hydrological cycle; Hydrology; Ice ages; Isotope composition; Isotope ratios; Land cover; land derived precipitation tracer; Ocean derived precipitation tracer; ocean‐sourced precipitation; Oxygen; Oxygen isotope ratio; Oxygen isotopes; Precipitation; precipitation oxygen isotope‐ratios; Rainout; Rayleigh distillation; Records; recycled precipitation; Recycling; Soil; Soil water; Tracers; Transpiration; Tropical climate; Water reuse; Water vapor; Water vapour
• ISSN: 0094-8276; 1944-8007
• DOI: 10.1029/2019GL084749

Large‐scale (>500 km) spatial gradients of precipitation oxygen isotope ratios (δ18Op) hold information about the hydrological cycle. They result from the interplay between rainout and evapotranspiration along air‐parcel paths, but these counteracting effects are difficult to disentangle, complicating quantification of the effect of land cover change on δ18Op. We show that disentangling can qualitatively be achieved using climate model simulations with a land‐derived precipitation tracer for tropical South America. We then either vary land cover as observed since 1870 or replace Amazon forests with bare land to determine the resulting signals. Our results indicate that effects of historically changing land cover on annual mean δ18O isotope‐ratio gradients are small and unlikely detectable, although there is a noticeable signal during the dry season. Furthermore, the effect of changes in water recycling on Amazon δ18Op in paleo‐records may have been overestimated and need reinterpretation. Plain Language Summary Deforestation causes reduction in precipitation downwind because trees act as pumps of water from soils to the atmosphere. This mechanism is primarily important during the dry season. How strong this effect is currently in the Amazon, given that approximately 20% of the forests have been cut, and how important it may be in the future if more forests are being destroyed is of great interest. One indicator of such changes is the east‐west difference in heavy water isotope content of precipitation. While preferential rainout of the heavy isotope along air parcel trajectories enhances this difference, transpiration by forests decreases the difference. This is because forests inject water back into the atmosphere that is more enriched than the overlying water vapor. Records of this difference during the last ice age, in particular, have been interpreted in a previous study as providing information on continental recycling. We apply a land‐derived water tagging approach in model simulations to investigate the effect of continental recycling on precipitation isotope content and to estimate this effect for varying land cover. We find that a 20% deforestation has only a small impact on precipitation isotope content. Even for a complete deforestation, in contrast to a previous interpretation, thus, only some of the isotopic signal observed during the ice age can be attributed to changes in continental recycling. Key Points We simulated Amazon δ18Op and contin