Switchgrass (Panicum virgatum L.) may have value as forage and a bioenergy feedstock. Our objective was to evaluate how harvest system and N fertilizer rates affected biomass yield and nutrient composition of young stands of switchgrass (cv. Alamo) in the southern Great Plains, USA. Nitrogen fertilization increased biomass yields from 10.4, 10.8, and 12.2 Mg ha −1 at 0 kg Nha −1 to 13.7, 14.6, and 21.0 Mg ha −1 at 225 kg Nha −1 when harvested after seed set (October), after frost (December), and twice per year after boot stage (July) and frost, respectively. Nutrient concentrations and removal were generally twice as great when biomass was harvested twice versus once per year. Precipitation strongly affected biomass yields across the two years of these experiments. When late-summer precipitation is available to support regrowth in this environment, harvesting switchgrass twice per year will result in greater biomass yields. Harvesting twice per year, however, will increase fertilization requirements and reduce feedstock biomass quality. Switchgrass harvested during mid-summer after boot stage was of poor forage quality. To have value as a dual-purpose forage and bioenergy feedstock, switchgrass would need to be utilized during spring to early summer while in a vegetative stage.

No abstract

Increasing desire for renewable energy sources has increased research on biomass energy crops in marginal areas with low potential for food and fiber crop production. In this study, experiments were established on low phosphorus (P) soils in southern Oklahoma, USA to determine switchgrass biomass yield, nutrient concentrations, and nutrient removal responses to P and nitrogen (N) fertilizer application. Four P rates (0, 15, 30, and 45 kg Pha −1 ) and two N fertilizer rates (0 and 135 kg Nha −1 ) were evaluated at two locations (Ardmore and Waurika) for 3 years. While P fertilization had no effect on yield at Ardmore, application of 45 kg Pha −1 increased yield at Waurika by 17% from 10.5 to 12.3 Mg ha −1 . Across P fertilizer rates, N fertilizer application increased yields every year at both locations. In Ardmore, non-N-fertilized switchgrass produced 3.9, 6.7, and 8.8 Mg ha −1 , and N-fertilized produced 6.6, 15.7, and 16.6 Mg ha −1 in 2008, 2009, and 2010, respectively. At Waurika, corresponding yields were 7.9, 8.4, and 12.2 Mg ha −1 and 10.0, 12.1, and 15.9 Mg ha −1 . Applying 45 kg Pha −1 increased biomass N, and P concentration and N, P, potassium, and magnesium removal at both locations. Increased removal of nutrients with N fertilization was due to both increased biomass and biomass nutrient concentrations. In soils of generally low fertility and low plant available P, application of P fertilizer at 45 kg Pha −1 was beneficial for increasing biomass yields. Addition of N fertilizer improves stand establishment and biomass production on low P sites.

Biomass demand for energy will lead to utilization of marginal, low fertility soil. Application of fertilizer to such soil may increase switchgrass (Panicum virgatum L.) biomass production. In this three-way factorial field experiment, biomass yield response to potassium (K) fertilizer (0 and 68 kgK ha −1) on nitrogen (N)-sufficient and N-deficient switchgrass (0 and 135 kgNha −1) was evaluated under two harvest systems. Harvest system included harvesting once per year after frost (December) and twice per year in summer (July) at boot stage and subsequent regrowth after frost. Under the one-cut system, there was no response to N or K only (13.4 Mgha −1) compared to no fertilizer (12.4 Mgha −1). Switchgrass receiving both N and K (14.6 Mgha −1) produced 18 % greater dry matter (DM) yield compared to no fertilizer check. Under the two-cut harvest system, N only (16.0 Mgha −1) or K only (14.1 Mgha −1) fertilizer produced similar DM to no fertilizer (15.1 Mgha −1). Switchgrass receiving both N and K in the two-cut system (19.2 Mgha −1) produced the greatest (P< 0.05) DM yield, which was 32 % greater than switchgrass receiving both N and K in the one-cut system. Nutrient removal (biomass×nutrient concentration) was greatest in plots receiving both N and K, and the two-cut system had greater nutrient removal than the one-cut system. Based on these results, harvesting only once during winter months reduces nutrient removal in harvested biomass and requires less inorganic fertilizer for sustained yields from year to year compared to two-cut system.

The parameters of yield response functions can vary by year. Past studies usually assume yield functions are nonstochastic or "limited" stochastic. In this study, we estimate rye-ryegrass yield functions where all parameters are random. Optimal nitrogen recommendations are calculated for two yield response functions: linear response plateau and Spillman-Mitscherlich. Nonstochastic models are rejected in favor of stochastic parameter models. However, the economic benefits of using fully stochastic models are small since optimal nitrogen rates do not differ greatly between stochastic and nonstochastic models. Previous work on crop response to nitrogen fertilizer has usually used either limiting nutrient response functions or polynomial models. Plateau functional forms tend to best fit data from field studies (Heady and Pesek 1954, Lanzer and Paris 1981, Grimm, Paris, and Williams 1987. Past studies have often assumed that the parameters of the yield function are nonstochastic or "limited" stochastic (some parameters are considered stochastic and others are not), and that all model errors are independent. This assumption often leads to estimating the parameter values of the assumed yield function by ordinary least squares. Research suggests, however, that parameters of yield response functions can vary by year.3 Random parameter models have been suggested by Berck and Helfand (1990), Paris (1992), Wallach (2002), andTembo et al (2008). Berck and Helfand (1990), and Paris (1992) consider linear response plateau models where the intercept and plateau parameters are random, but without random effects. Tembo et al (2008) adds uncorrelated random effects to the intercept and plateau, but not to the slope. Of these studies, only Makowski and Wallach (2002) treat all of the model parameters as random. Makowski and Wallach (2002) consider a linear-plus-plateau function in which wheat yield response is related to N uptake, and nitrogen uptake is related to applied nitrogen.We consider three crop response functions: the linear response plateau (LRP), the Spillman-Mitscherlich, and the quadratic; and we make all model parameters random. Our random parameter model lets parameters vary stochastically by year. The data used are annual rye-ryegrass forage data collected from a long-term nitrogen fertilization experiment in southcentral Oklahoma. We conduct nested likelihood ratio tests to choose between nonstochastic and stochastic models (Greene, 2008), and evaluate the economic value of using the alternative models by comparing expected profit. The ultimate goal is to make optimal nitrogen rate recommendations for cool season cereal rye (S.cereale)-ryegrass (Lolium multiflorum Lam)forage producers in southern Oklahoma. Determining the Profit Maximizing Level of Nitrogen FertilizerConsider a risk-neutral forage producer whose objective is to maximize expected net returns from winter cereal rye-ryegrass forage. The producer seeks to maximize expected net return above nitrogen cost:(1)where is the producer"s net return a...