Climate and weather risk in natural resource models

This work, consisting of three manuscripts, addresses natural resource management under risk due to variation in climate and weather. In three distinct but theoretically related applications, I quantify the role of natural resources in stabilizing economic outcomes. In Manuscript 1, we address policy designed to effect the risk of cyanobacteria blooms in a drinking water reservoir through watershed wide policy. Combining a hydrologic and economic model for a watershed in Rhode Island, we solve for the efficient allocation of best management practices (BMPs) on livestock pastures to meet a monthly risk-based as well as mean-based water quality objective. In order to solve for the efficient allocations of nutrient control effort, we optimize a probabilistically constrained integer-programming problem representing the choices made on each farm and the resultant conditions that support cyanobacteria blooms. In doing so, we employ a genetic algorithm (GA). We hypothesize that management based on controlling the upper tail of the probability distribution of phosphorus loading implies different efficient management actions as compared to controlling mean loading. We find a shift to more intense effort on fewer acres when a probabilistic objective is specified with cost savings of meeting risk levels of up to 25% over mean loading based policies. Additionally, we illustrate the relative cost effectiveness of various policies designed to meet this risk-based objective. Rainfall and the subsequent overland runoff is the source of transportation of nutrients to a receiving water body, with larger amounts of phosphorus moving in more intense rainfall events. We highlight the importance of this transportation mechanism by comparing policies under climate change scenarios, where the intensity of rainfall is projected to increase and the time series process of rainfall to change. In Manuscript 2, we introduce a new economic groundwater model that incorporates the gradual shift from irrigation to dryland farming as parts of an aquifer run dry. We accomplish this using an upside down cone to represent the spatial depletion, where the area of irrigable land above the aquifer shrinks as the water level decreases. Depletion of the aquifer may interact with uncertainty of the supply of water because the buffer that groundwater provides is no longer available. In this work, we identify the impact of spatial depletion on welfare gains from optimal management when rainfall is stochastic and follows a Markov process. Using a stylized model and dynamic programming, we estimate gains from moving away from current myopic extraction behavior to optimal use of the resource. When applied to Kansas over a section of the Ogallala Aquifer, we find gains from management ranging from 2.88% to 3.01% with larger gains achieved under uncertainty in the rainfall process. We find that including the dynamic of the gradual spatial depletion of the aquifer does materially impact welfare results compared to other estimates of the same region. Surprisingly the serial correlation of rainfall matters little. Empirically, multi-year droughts combined with the loss of access to the aquifer only slightly increases welfare gains due to the availability of dryland farming and the productivity of that option as a backstop when available. Manuscript 3 empirically estimates the effect of an increase in natural gas pipeline capacity in New England on monthly equilibrium natural gas prices and quantities for the electric sector. Weather plays an important role in defining the demand for natural gas due to its use for heating and electricity generation in the winter and through electricity demand for cooling in the summer. The cost of natural gas has important consequences to the wellbeing and cost of living for millions of customers either relying directly on natural gas for heating, or electric energy consumers indirectly. This paper presents results of reduced form price and quantity time series regressions using Generalized Least Squares (GLS) followed by results of a dynamic simultaneous equation model (SEM) of the market system. I highlight the role capacity has in effecting the variability of the price of energy to the region. This work adds to the literature by providing empirical evidence and the quantification of the effect of constrained pipeline supply in an important energy market, where weather conditions, multiple demand sectors and alternative fuels determine the cost of energy. I find that capacity is a significant factor in the prices and quantities of natural gas consumed by the electric sector, with an increase in pipeline capacity of 1% leading to an average decrease in price of .48% and an increase in consumption of .2%. The SEM model finds both supply and demand to be price inelastic. (Abstract shortened by UMI.