Simulating the dynamics of primary productivity of a Sonoran ecosystem: Model parameterization and validation
Modeling has played a crucial role in understanding the structural and functional dynamics of forest and grassland ecosystems in the past decades, but relatively few ecosystem models have been developed for deserts. Adapting an existing desert ecosystem model to new regions with different community components and environmental settings may testify to the generality of model applicability, further verify model structure and formulations, and provide new insight into understanding desert ecosystem functioning. In this paper, we use a desert ecosystem model that was originally developed for the Chihuahuan Desert, Patch Arid Land Simulator-Functional Types (PALS-FT), to estimate the aboveground annual net primary productivity (ANPP) of a creosotebush (Larrea tridentata)-dominated Sonoran Desert ecosystem in the Phoenix metropolitan area, home to the Central Arizona-Phoenix Long-Term Ecological Research Project (CAP LTER). We modified and parameterized the model using meteorological data, ecophysiological parameters for different plant functional types, and site characteristic data from the CAP LTER study area and an independent test site in the San Simon Valley of southeastern Arizona. Model predictions were validated and calibrated using field observations from the San Simon Valley test site. The results showed that PALS-FT was able to simulate ANPP of this typical Sonoran Desert ecosystem reasonably well, with a relative error of ±2.4% at the ecosystem level and generally less than ±25% at the functional-type level. We then used the model to simulate ANPP and its seasonal and inter-annual dynamics for a similar ecosystem in the CAP LTER study area. The model predicted average annual ANPP of 72.3 g m-2 y-1, ranging from 11.3 g m-2 y-1 to 229.6 g m-2 y -1 in a 15-year simulation. The simulated average ANPP of the Sonoran Desert ecosystem is close to field observations in other areas of the Sonoran Desert, and the range of variation also is close to that reported by other researchers for arid and semiarid ecosystems. The dynamics of ecosystem ANPP in response to fluctuations in annual precipitation simulated by the model agreed well with the known relationship between ANPP and precipitation in arid and semiarid systems. A closer examination of this relationship at the level of plant functional types further revealed that seasonal distribution of rainfall significantly affected ANPP. A comparison between the PALS-FT model prediction and two regression models for North American warm deserts showed that both regression models underestimated the Larrea ecosystem ANPP, while the process-based PALS-FT model provided the most accurate prediction among the three models. This study provides a validation for use of the PALS-FT model to investigate Sonoran desert ecosystem responses to environmental changes.
Shen, W; Wu, J; Kemp, PR; Reynolds, JF; Grimm, NB
Volume / Issue
Start / End Page
International Standard Serial Number (ISSN)
Digital Object Identifier (DOI)