As climate models become a primary tool for studying the atmospheric role of land surface processes, a question for current climate models is whether they can adequately distinguish and accurately simulate surface energy partitioning over different vegetation types. Plants contribute a large fraction of latent heat flux through evaporation of water from leaf surfaces and transpiration from deeper soil layers when stomata open during photosynthesis. Plants also affect net radiation by altering the surface albedo. A change in plant height can change the boundary layer turbulence by influencing surface roughness, and therefore the total energy exchange via latent and sensible heat fluxes [Davin and de Noblet-Ducoudre, 2010]. In most climate models, several important vegetation parameters are prescribed according to satellite observations and ground measurements. These parameters are not necessarily accurate at the site-scale due to the algorithm and validation methods used in retrieving satellite data or aggregating ground data [Yang et al., 2006]. Validation of surface fluxes over different vegetation types can help identify deficiencies in key parameters and model formulations to target for improving model performance. The aim of this work is to examine energy partitioning and surface climate simulated by a recently coupled regional climate model, WRF3-CLM3.5, for four major vegetation types across the United States, and to identify the model’s strengths and deficiencies to help prioritize model improvements. As the next-generation mesoscale numerical model,flower bucket the standard version of WRF includes relatively simple land surface schemes, which potentially constrain model applications for studying the land surface and ecosystem atmosphere feed backs.
The newly coupled model improved the surface process simulation in California [Subin et al., 2011], but has not been validated at the continental scale. We used the standard version of the Weather Research and Forecasting model version 3.0 [Skamarock et al., 2008], AmeriFlux site observations [Wofsy and Hollinger, 1998], and CERES data [Wielicki et al., 1996; Young et al., 1998] to evaluate energy flux partitioning. We analyzed the bias in surface climate variables by comparing to PRISM datasets [Di Luzio et al., 2008]. We focused on four dominant vegetation types with adequate representation in the AmeriFlux network . Both WRF and CLM have deficiencies that should be resolved in future versions to reduce the warm bias. The large warm bias in the standard version of WRF suggests there are problems in WRF. For example, the downward solar radiation bias contributes substantially to the warm bias based on a one-year sensitivity test that artificially reduced downward solar radiation by 30% . Reducing downward solar radiation is not simple because it is associated with many factors. Previous work [Markovic et al., 2008; Wild et al., 2001] suggests the overestimate of downward solar radiation at the surface could be either due to less cloud cover for cloudy days or less sky absorption of downward solar radiation for clear days. Ignoring aerosols in the model may also contribute to excess downward solar radiation [Wild, 2008]. The negative precipitation bias in the Midwest suggests that an underestimate of cloud cover may contribute to excess downward solar radiation in the Midwest. Further validation with WRF3-CLM3.5 focusing on the cloud cover and clear sky absorption are strongly encouraged but are beyond the scope of this paper. Fortunately, the newer WRF3.2 includes boundary layer physics and microphysics that could improve the overall simulation. A one-year sensitivity test using standard WRF3.2 with the MYNN boundary scheme [Nakanishi and Niino, 2009] and Thompson microphysics scheme [Thompson et al., 2008] showed a reduction in the downward solar radiation by 30 W m-2 , in T_max by 3K, and in T_min by 2K. With respect to CLM, the large warm bias in the Midwest could be due in part to 1) the missing irrigation scheme, and 2) the lower crop leaf area index used in the model.
A large area in the Midwest is covered by irrigated agriculture according to global irrigation maps [Siebert et al., 2005]. Without an irrigation scheme, WRF3-CLM3.5 may overestimate temperature by 3-5K in summer in the Midwest [Lobell et al., 2009; Sacks et al., 2009]. Considering the strong coupling between soil moisture and precipitation in the Midwest [Koster et al., 2004], low soil moisture could reduce cloud cover and enhance the downward solar radiation, further heating the land surface and contributing to a positive feedback in this region and producing a large warm bias. Also, the much lower maximum leaf area index used in the model could reduce LE and therefore increase H and near-surface temperature. The simulated seasonal variation in LAI is much lower than the direct measurements at the Bo1 site [Wilson and Meyers, 2007]. Mandatory marketing organizations have been an important agricultural policy tool in the United States for 80 years. These organizations include agricultural marketing orders, commissions, councils and check-off programs. They can serve many purposes, including supply control , setting of quality and grading standards, market or production research, limiting of unfair trade practices and generic commodity promotion. The intended purposes differ across MMOs and are outlined in each organization’s governing documents. The creation of a MMO is a political process, and considerable discretion is given to the Secretary of Agriculture in determining the value of these MMOs. In addition, many MMOs require approval through a vote of eligible producers for creation, continuation and termination. The outcome of this referendum then informs the Secretary’s decision for the future for the organization, and, in some cases, may dictate it. Producer referenda are held at regular intervals for most MMOs to ensure they are continuing to provide positive net benefits to producers. These referenda are the focus of our research. Specifically, we examine how market power among agricultural producers relates to voting power in a referendum to form, terminate or continue a MMO with a generic promotion provision. These referenda are interesting for several reasons. First, the voting rules used in determining the outcome of these referenda often depend on both the number of producers and the quantity of output they produce, suggesting that market structure matters in determining the outcome. Second, they provide an opportunity for us to study grower behavior regarding mandatory collective action organizations, which can shed light on the costs and benefits to growers of these organizations as well as on attitudes to collective action more broadly.
One of the most common types of MMOs is a marketing order. Although voting rules differ somewhat across MMOs, for the purposes of this paper, we consider voting rules for Federal marketing orders, as they are typical of the type of voting rule used by many MMOs. We model the supply side of the market for a homogeneous agricultural commodity as consisting of a single firm with a cost advantage and multiple firms with heterogeneous higher costs. This cost structure is intended to represent the supply environment in many industries,square flower bucket where there is an increasing gap between a few dominant producers and many smaller ones. We assume buyers of the commodity do not exercise market power. Finally, we focus on demand-increasing generic commodity promotion as the means by which an MMO benefits producers. Generic promotion is increasingly one of the primary roles of MMOs in the United States, in part due to the passage of the Research and Commodity Promotion Act of 1996, which created of a new category of federal check-off program with a major emphasis on generic promotion.1 We focus on a pair of voting rules commonly used together for Federal marketing orders and examine the voting power of the dominant and fringe firms. For this analysis, we consider what Felsenthal and Machover call “I-power.” This class of power measure address a voter’s influence over the decision to be made. The power measure we use is Banzhaf Power . The Banzhaf Power Index is calculated by considering all possible “winning” coalitions of voters—those coalitions that could pass a proposed action if all members of the coalition favored it, given the voting rule. A voter is “critical” if the coalition would no longer pass the proposed action if the voter left that coalition. The Banzhaf Power Index is defined as the number of times a given voter is critical out of the total number of possible vote combinations in the industry. Running simulations of these markets under various assumptions about costs and market structure and industry-calibrated market parameters, we calculate the Banzhaf Power Index value and market share of each firm. The Banzhaf Power Index assumes implicitly that voters vote for the action with a probability of 0.5. Some have challenged the usefulness of this type of measure given this naive assumption about voter behavior . However, developing a better model requires more information about voters, which can be hard to obtain. In most situations, voter preferences or correlations between preferences are difficult to measure and the factors that underly them may be challenging to identify.This voter characteristic allows us to incorporate our knowledge of theory of the firm to better predict the voter behavior given cost and market parameters.
Building on the probability theory approach to Banzhaf’s index identified by Straffin , and the behavior of profit-maximizing producers in the neoclassical theory of the firm to develop a second measure we call “Feasible” Banzhaf Power. To calculate this measure, we incorporate the information about each producer’s profit-maximizing voting choice and then assume that a producer votes in accordance with his profit-maximizing choice with a randomly drawn idiosyncratic probability. This probability represents the probability that a producer is optimizing some objective function other than profit-maximization that yields him or her to a different voting choice. For example, the producer could have the objective of minimizing government intervention, regardless of its effect on his profits. We find that market power and cost heterogeneity do indeed matter in determining the voting power of producers in MMO referenda, whether or not producer behavior is incorporated. Furthermore, incorporating information on producer behavior substantially affects our estimate of voting power. We also find that disparate preferences of firms with heterogeneous costs in situations where some producers wield market power can reduce the Feasible Banzhaf Power Index value for the low cost firm, even when the low cost firm produces a substantial share of industry output. Finally, we find that the different voting rules faced by producers in MMO referenda yield distinct differences in voting power in markets with heterogeneous producers. Our contributions to the literature are threefold. First, we contribute to both the voting power and agricultural economics literature, as to the best of our knowledge we are the first authors to examine marketing order referenda through the lens of voting power. Second, we contribute to the voting power literature by connecting the work on voting preferences and empirical voting power measures to the neoclassical theory of the firm in the form of our Feasible Banzhaf Power Index. This index is useful in that it incorporates information about behavior to provide a more realistic measure of voting power in settings with firms in the role of voter. And finally, through the analysis of voting power measures, we provide new insights about the potential challenges agricultural producers face in adapting their MMOs in a rapidly changing economic landscape. As agricultural market structures have changed over time, agricultural economists have moved away from the long-held assumption of perfect competition in some agricultural settings. Our work shows how the marriage of voting power methodology and agricultural economics can shed new light on how market power and cost heterogeneity interact with agricultural institutions in changing markets. The remainder of the paper proceeds as follows. In Section II, we give a brief history of MMOs and discuss the relevance of our work to agricultural policy. In Section III, we relate our work to the relevant literature in agricultural economics and political science. In Section IV, we present our theoretical model. In Section V, we discuss our simulations and calibration methodology. Section VI includes the presentation and discussion of our results, and Section VII concludes. MMOs were first authorized at the federal level by the Agricultural Adjustment Act of 1933 and the Agricultural Marketing Agreement Act of 1937. Initially, they were a policy response to ongoing low and volatile returns to agriculture in the 1920s and 1930s.