Cohort Level: Cohort - II

Career Goal: My career goal is to become a Data Scientist

Expected Graduation Date: May 20, 2019

Degree: M.S Economics

Research Title: An analysis of the Standardized Precipitation Evapotranspiration Index (SPEI) on drought based on Economic Impacts in the Northeast (US)

Research Synopsis: Title An analysis of the Standardized Precipitation Evapotranspiration Index (SPEI) on drought based on Economic Impacts in the Northeast (US) Introduction In order to examine causes of economic and/or environmental distress, it is important to understand drought. Drought is a byproduct of long periods without precipitation, but it is extremely difficult to quantify. Therefore, an index known as the Standardized Precipitation Index (SPI) has been used in previous studies to analyze droughts that are considered as months-years of below normal precipitation. This type of drought is known as meteorological drought. Absence of rainfall could be used as a tool to measure drought, but it does not consider evaporation; hotter and drier air can extract water from soil and plants. Thus, additional models can incorporate groundwater details or particular characteristics of plants. There are other types of droughts called Agricultural and Non-Agricultural. Agricultural drought is the reduction of plant growth and crop production due to below normal precipitation, less frequent rainfall and/or above normal evaporation which result in dry soil. Non-Agricultural drought on the other hand affects businesses, tourism etc. instead of crops. Using the SPI to evaluate/analyze these two types of drought would not be effective because this index is solely based on precipitation. For this reason, the Standardized Precipitation Evapotranspiration Index (SPEI) was created. This index was created to measure agricultural droughts, but it also takes into consideration, the meteorological droughts as well as crop evapotranspiration. In the Northeast (United States), many studies suggest that there are signs of drought occurrences. Therefore, in this study, we want to understand how “good” of a predictor the SPEI is by looking at the northeast to find trends between dryness indication of the SPEI and drought economic impacts. First, the SPEI was formatted to look at trends in drought-intensity-duration-frequency. Second, economic impact data is looked at to find out where there are trends in low production in both crops and business output. Problem analysis The Northeastern region consists of two divisions, New England and Middle Atlantic. The New England division consists of Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island and Vermont. The Middle Atlantic division has New York, New Jersey and Pennsylvania. In the northeast region, there have been evidence of global warming while the southeast has seen more of a cooling trend. Water availability and water resources are affected by both precipitation, and evapotranspiration. These are the inputs for water and the loss of water respectfully. The effects of drought on the Northeastern region includes but are not limited to agriculture, Fisheries and water resources. About 175 thousand farms are located in the Northeast that produce above $21 billion goods per year. In 2016, many producers lost on crop yields, hay and corn production also suffered which had a correlation with dairy farmers and grazed animals. Top soil has seen its driest yet since the late 1990’s. Since these conditions were not affecting the entire country, commodity prices were driven to a multi-year low. As for fisheries, decrease of stream flow caused a negative impact on the fish species in the New England rivers. Many fish deaths were observed due to the lack of stream flow. With water resources, major portions of New York and New England areas experienced below-normal precipitation levels and as a response, waterways received extremely low stream flows. The Catskill reservoir in upstate New York that feed New York City has seen about a 20% capacity drop in the overall system functionality. The challenges of drought include farmers having a lack of irrigation equipment, crop insurance, accurate long-range forecast, and access to data and products that could lead to better crop management. Another challenge is that the monitoring of drought has suffered due to loss of several U.S.G.S. wells as well as the failure of existing and older wells used to monitor groundwater. There is also the problem of conflicting drought condition information between U.S. drought monitoring, the groundwater level observations and the locally collected data which leads to confusion among decision makers and the public. Early drought predictions are considerably important as it pertains to observations of private wells, however the Northeast region lacks the inventory of existing private wells or real-time impact surveys. Lastly, a challenge of drought is the lack of information on the impact that low flows have on recreational use of rivers. Objectives Given the SPEI, the user will be able to ensure that the frequency of extreme events will be consistent on any time scale due to its standardization. The SPI is calculated by fitting probability density function to the frequency distribution of the precipitation that is summed over the timescale chosen. To define meteorological drought, the SPI is a valuable tool for the estimation of drought intensity, frequency and duration. In terms of intensity, frequency and duration of drought, the SPI method is able to return fundamental parameters in the analysis of drought. Given the SPEI data, drought will be defined by precipitation and temperature. For instance, a SPEI value of 2 or more can be considered as extremely wet and a SPEI or -2 or less can be considered as extreme drought as seen in table 1. To measure the duration and intensity of droughts, we will use the SPEI to quantify the duration. For instance, we will try to make a correlation between the extreme droughts and duration. It may be the case that where there are extreme droughts, their durations are shorter and vice versa. Methodology and methods 3.1. SPEI drought duration-intensity-frequency One of the most efficient ways of dealing with the SPEI data is to reformat it as a data frame within Rstudio. The data is in a netCDF format with three dimensions. The dimensions are longitude, latitude and time. Since the structure is spei[lon,lat,time], to do any data exploration, the first two dimensions which are longitude and latitude were extracted, then for all those grid cells were SPEI values from 1901 to 2015. Each year has 12 months. To break into the SPEI, looking at the duration, intensity and frequency is important. Given that the SPEI12 is always based on the previous 12 months, the December month would be based on all the months in that calendar year. Therefore, SPEI12 will be used primarily to analyze the duration, intensity and frequency. First, as said previously, the longitude (lon), latitude (lat) and time variables are all pull out separately from the unstructured spei data. Once that is done, they are all attached back together into a dataframe where the lon and lat are the first two columns. The third to the last column represents the time. What this shows is an spei value for each grid cell in a particular month. Since the grid cells are ranged widely, the locations need to be restricted to the northeast locations. Now, the drought duration, intensity and frequency can be found. Each drought will last for a certain number of months (duration), each drought period will have a certain average SPEI value (intensity), and we can test whether the durations or intensities have been increasing or decreasing. Since each location will have a different length, we will need to run loops and check for the p-value and slopes. 3.2. Drought economic impacts/Agricultural sector Common signs of drought in the agricultural sector is dried crops, yellow pastures and abandoned farmlands. This is due to the prolonged deficit of soil moisture that causes dried crops. Drought losses are not solely to farmers as they are as well to the consumers when the prices increases. Temperature and precipitation changes have a direct impact on agricultural production. Different studies have focused on analyzing the impacts of climate changes on agriculture at different geographical scales however, not many studies have focused on using standardized precipitation evapotranspiration index (SPEI) to estimate a timescale that can be used when quantifying drought. In this study we will look at crops that are grown in the northeast like Barley, Corn, and Wheat. The data for these crops is taken from the United States Department of Agriculture (USDA). Depending on the time frame of the data, we will be looking a time where there is economic prosperity/no economic hardship and drought that is prevalent. With this is mind, we will be looking at crop acre-harvested, crop acre-produced, and crop acre- planted for three different crops (Barley, Corn and Wheat). 3.3. Drought economic impacts/Non-Agricultural sector Non-Agricultural drought sectors include Effect on water supply (streamflows, reservoirs, wetlands, and groundwater). Addition Sectors are tourism, recreation, public utilities horticulture and landscape services, navigation and other industries /businesses that have significant water consumption. It is common to equate drought losses to farmers losses but, that is incorrect because farmers may have farm insurance or the government may aid in disaster relief. The economics impacts on drought can be very difficult to assess because, the drought induced highers prices will allow and inflow of producers outside the drought stricken area that are benefiting from the higher prices. Therefore It is important to set a time frame because drought impacts are long term as well as short term local drought impacts might be cancelled out when evaluated at regional or national level zero-sum transfers of losses or gains should be excluded from impact assessment (Griffin 1998)