Tree Water Use Model Validation

Graphic of tree enclosed with three sections of plastic    bags to isolate and control environmental conditions within the canopy

 

The limitations of direct measurement techniques (e.g. expense, small scale) emphasize the need to go beyond whole-tree measurements of sap flow into measuring stands or blocks of trees, a necessary component of nursery scale water use modeling.

 

Modeling is a valid and viable means to estimate water use and scale individual tree estimates to groups of trees on a nursery. To date, simplified forms of the Penman-Monteith equation (the universally accepted engineering equation for potential evapotranspiration calculation) are often used to estimate rates of canopy transpiration.


Processed Based Evapotranspiration Models

measured vs. modeled transpiration rate by soil moisture  availability over time

The Colorado State Predictive Modeling Group has taken the physiological and physical parameters that we collect in our experimentation and developed process-based models.

To validate and improve upon the models, we compare the estimated transpiration values from our models with observed values measured by various techniques. 

For example, sap flow gauge measurements of tree transpiration are compared to model estimates of transpiration (Figure 1). We compare these estimated and observed values of transpiration on a 15-minute, hourly, and daily basis.

Overall, the current whole tree model has an approximate 9% error in comparison to the measured tree transpiration rates.  This error percentage does not appear to change in relationship to the environmental conditions (e.g., sunny and cloud free days versus rainy and cloudy days).

Nursery Application

whole tree monitoring system in a nursery

The whole tree model can be tested at nurseries, where numerous trees are monitored under various environmental conditions. Experimental manipulation allows us to manipulate environmental conditions and force a wide range of potential conditions.  By comparing the model estimates with measured values over a wide range of environemtnal conditions, we can pinpoint where the models are most likely to fail and identify both their weaknesses and strengths.  Then, we can refine the models to achieve a high degree of accuracy under dynamic growth and environmental conditions.

Mechanistic models are a powerful and robust means to estimate tree water use in both nurseries and landscape trees.  Once perfected and validated, the models provide a realistic and economically feasible method of quantifying tree water use under almost any imaginable nursery or landscape condition.  Furthermore, the models we developed are capable of estimating values on a genotype-by-genotype basis (this component is very important given the amount of cultivars and species in the ornamental nursery trade).

For further information or for any questions please contact Dr. Bill Bauerle at bauerle@colostate.edu.