This study introduces a new analytical procedure based on the chi-square (x²) statistic to evaluate tree- ring weather relationships. An iterative x² method, developed previously for relating annual crop production to daily values of meteorological measurements, is applied to tree-ring data and compared to results obtained from correlation and bootstrapped response function analyses. All three analytical procedures use a southern Arizona chronology (Pinus arizonica Engelm.) and the latter two use monthly average meteorological data. The x² analysis revealed most of the relationships exhibited by the correlation and response function analyses as well as new linear and nonlinear associations. In addition, cardinal values were obtained that define daily thresholds of the meteorological variables at which the limitation to growth becomes significant. Some of the associations are plausible from the physical system but require more study to confirm or refute a real cause and effect. A few associations appear to be too late in the season or too early in the previous year to affect ring width. We recommend that this x² technique be added to the existing dendroclimatic procedures because it reveals many more possible cause and effect relationships.

Climate and longleaf pine (Pinus palustris P. Mill.) radial growth relationships have been documented within its southern and western distribution. However, knowledge of this relationship is lacking along its northern latitudinal range margin (NLRM). Based on the principles of ecological amplitude, limiting factors, and studies of coniferous species in eastern temperate forests of the U.S., we hypothesized that the radial growth of longleaf pine in mixed pine-hardwood forests is responding to winter temperatures in southeastern Virginia. Two longleaf pine chronologies were developed to determine the relationship between radial growth and monthly temperature, precipitation, and Palmer Drought Severity Index (PDSI) via response function analysis (RFA). Results at the 0.05 level yielded significant response function coefficients with a positive response to current winter temperature and precipitation and a negative response to prior August PDSI. In studies of climate and longleaf pine radial growth in other parts of its range, winter temperature and precipitation have not shared a significant positive association with radial growth. Instead current spring and summer precipitation usually share this positive association. These findings add more evidence to an emerging pattern suggesting that winter temperatures contribute to limiting the radial growth of temperate conifers at northern range margins in the Northern Hemisphere.

Although the primary purpose of response function analysis is to identify climate variables that have significant associations with tree radial growth, many researchers are also interested in assessing the strength of these associations. Existing response function programs use a liberal criterion to determine how many climate variables should be included in the analysis. The resulting response function models include a large number of predictor variables. The objective of this analysis is to determine if these response function models are over-fitted to the data used to calibrate them, resulting in over-estimation of strength of associations. PRECON was used to produce response functions for white oak chronologies from n = 149 sites, with separate response functions using 34 monthly climate variables or 10 seasonal climate variables. An analysis of goodness-of-fit statistics for response function calibration provided strong evidence of over-estimation of strength of associations. The degree of over-estimation was greater when 34 monthly climate variables were included in the models compared to models with10 season variables. There was much less evidence of over-fitting for the R-verif statistic that reflects strength of association between predicted and actual tree-ring indices that were not included in model calibration. The PRECON R-verif statistic is the best measure of the strength of multivariate growth-climate associations currently available.

South Florida slash pine (Pinus elliottii var. densa) is the southernmost pine species in the United States and the foundation species of the globally endangered pine rockland communities in south Florida. To test if slash pine produces annual growth rings in the Lower Florida Keys, we counted the number of rings on samples collected from the North Big Pine Key site (NBP), which contained a fire scar from a known wildfire and a known date for hurricane-induced tree mortality (2006 or 2007). In addition, a crossdated tree-ring chronology (1871–2009) was developed from living trees and remnant wood found at the site and compared to divisional climate data to determine how the regional climate regime influences radial growth. Our analyses demonstrated that slash pine forms anatomically distinct, annual growth rings with the consistent year-to-year variability necessary for rigorous dendrochronological studies. Response-function and correlation analysis showed that annual growth of slash pine at NBP is primarily influenced by water availability during the growing season. However, no significant correlations were found between tree growth and the Atlantic Multidecadal Oscillation or the El Niño-Southern Oscillation. Our study reveals the potential of producing high-quality dendrochronological data in southern Florida from slash pine, which should prove useful in further studies on fire history and tree phenology and for assessing the projected impacts of impending climate change on the fragile pine rockland community.

Doctor of Philosophy / Department of Economics / Wayne Nafziger / This dissertation constructs three empirical essays. The first essay illustrates the causality on the relationship between output (GDP) growth and exports. By using the Modified Wald (MWald) test we observe unidirectional causality from exports to GDP. More specifically, for the robustness we use a Vector Error Correction Model (VECM) model and the Generalized Impulse Response Function Analysis (GIRA). The VECM and the GIRA yield bidirectional causality between exports and GDP, which weakly supports the unidirectional result of the to MWald test. Meanwhile, we confirm that there is structure break by using the structural break test. These results are plausible and consistent with the expectations of our study for the Export Led Growth Hypothesis (ELGH). However, compared with previous studies on the ELGH for Korea, our results are different. Other studies show a bidirectional causality relationship but this study only has unidirectional causality. These differences may be caused from different observation data, various variables, and use of different econometric methodologies. Also, model selection and omitting variables can also significantly change the results of causality testing. The second essay investigates a degree of competition between Korea's and China's exports in the U.S. market by using the substitute elasticity on a simple demand model. The market share of Korean exports has been decreasing while that of China's has been increasing. The results of this study are as follows. First, we find that Korea has a dominant market share of only goods group code 27 in commodity groups over that of China, otherwise having China's dominant market shares over those of Korea for other export sections by using historical trade data. Second, most estimates of substitute elasticity between both countries' exports in the U.S. market are small (inelastic). However, 61 (apparel articles and accessories, knit or crochet), 62 (apparel articles and accessories, not knit etc) and 85 (electric machinery etc, sound equipments, TV equipment, parts) commodity groups' substitute elasticities are large (elastic) and are competitive in the U.S. market compared with those of China. A small value of the elasticity of substitution may be due to an identification problem for a simple standard model as well as measurement errors in prices as a unit value in this study. So, in order to avoid problems such as these, we may need to use appropriate instrumental or proxy variables in the simple standard model, which highly correlate with the independent (unit price) variables and are uncorrelated with measurement error terms. In practice, it is not easy to find good instrumental variables. The final essay evaluates the roles of price and income as important factors that affect Korea's exports by using the most recent monthly data. By using the Autoregressive Distributed Lag (ARDL) bounds testing approach we find the long-run relationship of variables and estimate the long-run price and income elasticities. However, the estimates of these long-run elasticities are statistically insignificant. This may be due to some misspecifications or measurement errors in our model. Meanwhile, due to the existence of the long-run relationship between variables, we construct the Error Correction Model (ECM) in order to observe the short-run dynamics of the elasticities. Specifically, we add a dummy variable into our export demand model to achieve more efficient estimations since the dummy variable reflects a shock in Korea's export; Korea's economic crisis in 1997. In contrast to the long-run elasticity, we find that the short-run elasticities' estimates are more statistically significant. When we use the structure break test to check the structural stability of Korea's export demand, we find that there is no structural break point of 1997. Therefore, a shock of Korea's economic crisis in 1997 might not significantly affect Korea's export demand in a given sample. However, the Information Technology (IT) bubble of the world economy in 2001 and the entry of Korea into the OECD had triggered an increase in Korea's export demand due to existing structural break points of both events. In addition, we find that income elasticities are larger than price elasticities in the short run. This implies that income has more of an impact than that of price for the export demand model in the short run. This also implies that the change of Korea's exports in the short run is more sensitive to changes in foreign income (industrial production) compared with that of price (exchange rate). An interesting result, thus, is that Korea's exports in the short run may have higher export performance on income than that of price (exchange rate). This might be a consequence of the dependence of an increase in foreign income in recent years. In recent years, developing countries have greatly increased their economic growth compared with that of developed countries and Korea's exports have increased into these developing countries. Thus, we confirm that an increase in Korea's exports is mainly affected by income compared with price, specifically in the short run by using recent data.