PS 38-53 - Design of a Field Experiment to Disentangle Soil Moisture and Vapor Pressure Deficit as Determinants of Net Primary Productivity

Abstract: Net primary productivity (NPP) can be influenced by both soil moisture and the atmospheric Vapor Pressure Deficit (VPD), but their relationships are usually interdependent. With climate change altering precipitation (PPT) regimes that determine soil moisture as well as altering VPD levels, understanding and disentangling the interactions between VPD, PPT, and NPP has become increasingly important. This is particularly true for water-limited ecosystems where soil moisture has historically been assumed to be the dominant driver of the carbon cycle.



Field experiments represent a critical element of understanding ecological functions such as VPD and PPT, but to date, nearly every VPD study relies on observational data or greenhouse experiments, neither of which adequately describe the impacts of these abiotic drivers on natural ecosystems. We contend that a new generation of field experiments are needed to modify both VPD and PPT independently rather than relying on observing existing VPD levels. Experimental infrastructure such as the ARMS (Automated Rainfall Manipulation System) has been in use for several years effectively modifying PPT inputs and soil moisture levels, and although there have been past attempts to modify VPD, there are very few tested experimental designs for assessing VPD x soil moisture interactions. Here we compare various methodologies for altering VPD and soil moisture under field conditions and recommend the most feasible and effective setup for VPD field experiments.



Altering VPD necessitates modifying one of two variables: humidity or temperature. Whichever variable is selected for modification must then ensure that the other (humidity or temperature) remains unchanged for to accurately compare VPD measurements. Given that warming experiments are prone to modifying humidity levels, we elected to modify the humidity and maintain constant temperature.



In a desert grassland we designed field infrastructure with 16 misting heads evenly spaced around 2.5m x 2.5m plots with each head 60 cm above the ground and capable of emitting ½ gallon of water per hour. Plant cover included local grass and forb species. To avoid any additions to soil moisture, various misting timing patterns were tested alongside control plots. Relative humidity and temperature were sampled every ten seconds to calculate VPD. We were able to alter VPD from .5647 kPa to 1.6094 kPa relative to ambient conditions depending on the temporal frequency of misting pulses applied. We conclude that it is possible to manipulate VPD and PPT to understand the independent and interactive effects of these variables on NPP in the field.