The genetic variation among ecotypes makes switchgrass challenging to model. The environmental and developmental cues that effect green up, leaf area development, and flowering time need to be well understood for each ecotype. Despite numerous field studies with many of these genetic lines at different locations, isolating the relationship between temperature, drought, and nutrient stress on green up, biomass accumulation, flowering time is difficult. This is because of variation in site-specific factors such as current management, previous management, soil type, and microclimate.

There have been two approaches taken to model biomass production of different ecotypes or groups of ecotypes. The ALMANAC model was used to simulate biomass of four groups of ecotypes (northern lowland, northern upland, southern lowland, and southern upland) at five locations (Kiniry et al. 2008). The model was parameterized with a different maximum leaf area index (LAI) for each ecotype based on how well adapted it was to the location’s climate and environmental conditions. Growing degree days also varied among groups of ecotypes. The maximum potential LAI was assumed to be larger in southern regions and largest for southern lowland ecotypes. Modeled versus measured yields at these locations show that the ALMANAC model did reasonably well at predicting yield all four types at the locations (Table 2). The model had its greatest errors when estimating biomass of the northern lowland type in KS and NE. Yields for the southern upland type were slightly overestimated at each location.

The DAYCENT model has also been used to simulate different switchgrass ecotypes; in this case two lowland ecotypes, Alamo and Kanlow, and four upland ecotypes, Blackwell, Cave-in-Rock, Sunburst, and Trailblazer (Lee et al. 2011). Simulations were validated for four locations in Central Valley of CA. It was assumed that each cultivar had the same

Field trial data from Casler et al. (2004). Table reproduced from Kiniry et al. (2008).

maximum LAI. Instead, the root to shoot ratio, baseline temperature, optimum temperature, and maximum temperature were adjusted for each cultivar. The upland types (Blackwell, Cave-in-Rock, Sunburst, Trailblazer) were assumed to allocate more primary production to root biomass than the lowland (Alamo and Kanlow) types. The lowland ecotypes were also assumed to have a higher optimum and maximum temperature. The DAYCENT model also did a reasonable job estimating the observed yields with the R2 of the one-to-one relationship between observed and measured ranging from 0.66 to 0.90.

Modeling Water Use Efficiency

Water use efficiency (WUE) can be vitally important for predicting which areas are suitable for switchgrass biofuel production. To avoid competition with farmland already used for food, fiber, and feed production, the areas most likely for switchgrass production are on less productive soils where soil, water and nutrients often limit production (Perlack et al. 2005). In dryland production systems, limited rainfall and/or limited capacity of soils to store moisture becomes an important issue for switchgrass production. Plant water use becomes a major concern in irrigated regions, where competition between agriculture and other water demands arise. Direct measurements of WUE are important but require labor-intensive procedures involving soil water measurement with neutron access tubes, gravimetric measurements of soil moisture from soil cores, or use of weighing lysimeters. Likewise, measurements of WUE require plant harvesting to determine dry weight of plants. To adequately define WUE for a range of soils, plant species, and climatic conditions will require considerable resources and time.