Geometric consistency for regression model estimation and testing in climatology and meteorology

Abstract

The effects of outliers on linear regression are examined. The sensitivity of classical least‐squares (LS) procedures to outliers is shown to be associated with the geometric inconsistency between the data space and the analysis space. This is illustrated for both estimation and inference. A geometrically consistent procedure based on the Euclidean distance is proposed. This procedure involves the least absolute deviation (LAD) regression and a new permutation test for matched pairs (PTMP). Comparisons made with LS techniques demonstrate that the proposed procedure is more resistant to the existence of outliers in the data set and leads to more intuitive results. Applications and illustrations using meteorological and climatological data are also discussed.     

Evolution of the Parisian urban climate under a global changing climate

The evolution of the Parisian urban climate under a changing climate is analyzed from long-term offline numerical integrations including a specific urban parameterization. This system is forced by meteorological conditions based on present-climate reanalyses (1970–2007), and climate projections (2071–2099) provided by global climate model simulations following two emission scenarios (A1B and A2). This study aims at quantifying the impact of climate change on air temperature within the city and in the surroundings. A systematic increase of 2-meter air temperature is found. In average according to the two scenarios, it reaches +?2.0/2.4°C in winter and +?3.5/5.0°C in summer for the minimum and maximum daily temperatures, respectively. During summer, the warming trend is more pronounced in the surrounding countryside than in Paris and suburbs due to the soil dryness. As a result, a substantial decrease of the strong urban heat islands is noted at nighttime, and numerous events with negative urban heat islands appear at daytime. Finally, a 30% decrease of the heating degree days is quantified in winter between present and future climates. Inversely, the summertime cooling degree days significantly increase in future climate whereas they are negligible in present climate. However, in terms of accumulated degree days, the increase of the demand in cooling remains smaller than the decrease of the demand in heating.     

An FDTD method for analysis of scattering from rough fluid-fluidinterfaces

A finite-difference time-domain (FDTD) method for scattering by one-dimensional, rough fluid-fluid interfaces is presented, modifications to the traditional FDTD algorithm are implemented which yield greater accuracy at lower computational cost. These modifications include use of a conformal technique, in which the grid conforms locally to the interface, and a correction for the numerical dispersion inherent to the FDTD algorithm, Numerical results are presented for fluid-fluid cases modeling water-sediment interfaces. Two different roughness spectra, the single-scale Gaussian roughness spectrum and a multiscale modified power-law spectrum, are used. The Gaussian results are calculated as a function of the dimensionless parameters kh and kl, where k is the wavenumber in water, h is the rms surface height, and l is the surface correlation length. For the modified power-law spectrum, statistical parameters consistent with an insonification frequency of 7.5 kHz are used. Results are compared with those obtained using an integral equation technique both for scattering from single-surface realizations and for Monte Carlo averages of scattering from an ensemble of surface realizations. Scattering strengths are calculated as a function of scattering angle for an incident angle of 70° (20° grazing). The results agree well over all scattering angles for the cases examined     

Assessing selected natural and anthropogenic impacts on freshwater lens morphology on small barrier Islands: Dog Island and St. George Island, Florida, USA

Abstract The freshwater lens morphologies of the barrier islands Dog Island and St. George Island on the panhandle coast of Florida (FL), USA, are controlled to varying degrees by both natural and anthropogenic factors. Variable-density groundwater flow models confirm that spatial variability of recharge values can account for the observed lens asymmetry on these islands. The depth to the base of the lens does not vary significantly seasonally. Human development has altered recharge patterns in some areas, locally thinning the freshwater lens. Aqueduct water supply to St. George Island represents 7–25% of natural recharge; higher recharge rates are required to simulate the lens on St. George Island than on Dog Island. On both islands, coastal erosion rates are sufficiently rapid that the freshwater lens may not be in equilibrium with current boundary conditions.

---

Resumen Las morfologías de los lentes de agua dulce en las islas de barrera: Islas Dog y St. George, en la costa alargada y estrecha de Florida, EUA, están controladas en diferentes grados por factores antropogénicos y naturales. Los modelos de flujo con densidad variable del agua subterránea, confirman que la variabilidad espacial de los valores de recarga, pueden ser la causa de la asimetría observada en los lentes de estas islas. La profundidad hasta la base de los lentes no varía significativamente con las estaciones climáticas. El desarrollo humano ha alterado las tendencias de recarga en algunas áreas, causando adelgazamiento local de los lentes de agua dulce. El abastecimiento de agua del acueducto en la Isla St. George, representa entre 7% y 25% de la recarga natural; por otro lado se necesitan cantidades mayores de recarga para simular los lentes en la Isla St. George que en la Isla Dog. En ambas islas, las tasas de erosión costera son tan rápidas, que las lentes de agua dulce podrían no estar en equilibrio con las condiciones actuales de frontera.

---

Résumé La morphologie des lentilles deau douce des îles de Dog et de St. George, situés dans le district côtier de Florida-Etats Unis est contrôlé dans une mesure variable par des facteurs naturelles et humaines. Les modèles découlement à densité variable des eaux souterraines confirment que la variation spatiale de la recharge peut expliquer les asymétries observées des lentilles des cettes îles. La profondeur des lentilles ne présente pas des variations saisonnières importantes. Les activités humaines ont modifié la recharge dans certaines zones en amincissant localement les lentilles deau douce. Lalimentation en eaux de lîle de St. George représente 7–25% de la recharge naturelle; pour simuler les lentilles de lîle de Dog il a été nécessaire dintroduire des valeurs plus élevées de la recharge par rapport aux ceux utilisés pour lîle de St. George. Pour les deux îles lérosion côtière est assez rapide ainsi que les lentilles de eau douce ne sont pas en équilibre avec les conditions à la frontière.