Forest Dormancy and LAI

[Aashi Agarwal]

I am working on simulating deciduous forest using the SWAT model. However, the default forest configuration does not accurately replicate the observed or remotely sensed LAI time series. Specifically, it exhibits an abrupt decline in LAI to 0, which is inconsistent with actual forest behavior.

To address this, I attempted several adjustments, including modifying the LAI shape parameters (LAIMX1, LAIMX2, FRGRW1, FRGRW2, DLAI, BLAI), heat unit settings, and management operations such as IGRO, Plant Only, and Plant + Kill. These changes improved the LAI simulation to some extent, but the sharp drop in 17-18th December persists.

Given that I have set IDC = 7 (Trees) and the latitude is 20°N, dormancy is triggered solely by the minimum daylength threshold. According to the literature, dormancy and the Kill operation are the only mechanisms in SWAT to reset the forest growth cycle. However, as per my understanding using the Kill operation for forests is not recommended, as it eliminates all biomass.

At this point, I'm unsure what changes should be made in the SWAT source code to delay the onset of dormancy or to prevent this unrealistic LAI decline.

I also experimented with simulating the forest as a warm-season annual to bypass dormancy, which resulted in a LAI drop during February instead. While this approach yielded a smoother LAI curve, I’m uncertain whether it is appropriate or realistic for forest ecosystems.

I have attached the simulated LAI outputs for both the forest and warm-season annual configurations for reference. 

I would appreciate any guidance on how to better simulate realistic LAI seasonality for deciduous forests in SWAT, especially in terms of modifying dormancy behavior.

Kindly help.