Paper Review: Guo (2019). SAM->iso/anisohydry. NPH

20 May, 2020 (Guo, NPH, SAM->iso/ansiohydry)

Guo, J.S., Hultine, K.R., Koch, G.W., Kropp, H. and Ogle, K. (2020), Temporal shifts in iso/anisohydry revealed from daily observations of plant water potential in a dominant desert shrub. New Phytol, 225: 713-726. doi:10.1111/nph.16196

A key reference for applying the SAM framework

I have reviewed a paper in which the SAM framework was applied to fluxtower NEE data. I read the paper because I used a similar type of data (i.e. fluxtower GPP and ET). However, after looking into the source code by Liu et al. (2019), I have found that it was more complicated than one I want to implement.

Instead, I found this paper. Guo et al. provides an excellent reference case with relevant R & JAGS source code of sufficiently complex design.

Conclusions about the advantages of iso/anisohydry from this paper are not that unique: isohydry helps trees to exploit the favorite environmental condition, whereas anisohydry makes them to survive during the dry condition. However, the authors could derive which environmental drivers affect those responses thanks to the SAM framework.

SAM framework quantifies the effects of environmental drivers to parameters in a model.

It was interesting that the SAM framework can be used to quantify the weight of effectiveness of drivers to model parameters as well as to a dependent variable. The authors could estimate time-varying parameter values and the relevant weights. Also, the framwork could be applied to quantify effects of drivers across various spatio-temporal scales, which could probably a key technique of my research.

Via inspecting the resulted weights of environmental drivers, they confirmed that the shife between iso/anisohydry was influenced immediately by soil water and slowly by temperature. Also, they found that the shift from isohydry to anisohydry in winter was caused by low temperature, not by low moisture content, referring highly negative weight of temperature during that time.

The Martinez-Vilalta framework

The MV framework is simple, but surprisingly outperformed to simulate midday water potential.

It is a simple linear regression:

MD = sig x PD + gam

where MD is midday water potential, PD is predawn water potential, and sig and gam are the slope and intercept, respectively.

The sig varies from 0 (isohydry) to higher values (anisohydry).

The MV framework with time-varying parameters could provide better estimates of the MD then the model with time-static ones.

This study described that the Larrea could successfully survive from dried region as it can flexibly shift between iso/anisohydry along with the environmental condition. They could be an anisohydry when the conditions are good, while they turned to isohydry with a harsh condition.

This may be a similar message with that a tree species can succeed in surviving during drought with flexibly adjust the water uptaking depth.