Impacts of localisation in the EnKF and EnOI: experiments with a small model

The performance of an inexpensive, ensemble-based optimal interpolation (EnOI) scheme that uses a stationary ensemble of model anomalies to approximate forecast error covariances, is compared with that of an ensemble Kalman filter (EnKF). The model to which the methods are applied is a pair of “perfect”, one-dimensional, linear advection equations for two related variables. While EnOI is sub-optimal, it can give results that are comparable to those of the EnKF. The computational cost of EnOI is typically about times less than that of EnKF, where is the ensemble size. We suggest that EnOI may provide a practical and cost-effective alternative to the EnKF for some applications where computational cost is a limiting factor. We demonstrate that when the ensemble size is smaller than the dimension of the model’s sub-space, both the EnKF and EnOI may require localisation around each observation to eliminate effects of sampling error and to increase the effective number of independent ensemble members used to construct an analysis. However, localisation can degrade an analysis if the length-scales of the localising function are too short. We demonstrate that, as the length-scale of the localising function is decreased, localisation can significantly compromise the model’s dynamical balances. We also find that localisation artificially amplifies high frequencies for applications of the EnKF. Based on our experiments, for applications where localisation is necessary, the length-scales of the localisation should be larger than the decorrelation length-scales of the variables being updated.     

Percolation losses in paddy fields with a dynamic soil structure: model development and applications

The hydraulic characteristics of the plough pan of paddy fields provide continuous ponding conditions during the growing season and control the water use efficiency in wet rice production. Its saturated hydraulic conductivity K_s, however, exhibits a large spatiotemporal variability as a consequence of a highly dynamic soil structure involving temporary shrinkage cracks. Water flow through the earthen bunds surrounding the fields further contributes to the uncertainty in water flux calculations. The objective of this study was to develop a simple deterministic model with stochastic elements (‘PADDY‐FLUX’) for depiction of deep percolation, and to assess the effect of different water management scenarios on percolation in two channel command areas. Darcy's law is used as the fundamental equation for water flow calculations with the ponding water depth h as a time‐dependent variable. Flux uncertainty is estimated by a Monte‐Carlo‐type implementation. K_s is treated as a random variable of a bimodal probability density function (PDF), which is the weighted sum of two Gaussian PDFs (accounting for a matrix and a preferential flow domain). The weighing factor α is a function of h, reflecting an increasing risk for preferential flow situations after desiccation and the development of shrinkage cracks. Under‐bund percolation is calculated using transfer functions. The results demonstrate that percolation losses increase in the following order: continuous soil saturation < continuous flooding (CF) < mid‐season drainage and intermittent irrigation (MD + II) < mid‐season drainage and continuous flooding. The bunds contribute up to 54 and 17% to total fluxes under CF and MD + II, respectively. Preferential water fluxes are responsible for the major part of water losses as soon as desiccation causes the formation of shrinkage cracks. As a conclusion, continuous soil saturation should be promoted as the least water‐intensive irrigation regime, while intermittent irrigation is recommended only in case that irreversible shrinkage cracks have already developed. Copyright © 2010 John Wiley & Sons, Ltd.     

Point source identification using a simple permutation test: a case study of elevated PCDD/F levels in ambient air and soil and their relation to the distance to a local municipal solid waste incinerator

A point source for environmental pollution may be identified by the spatial association of samples collected in the vicinity of the emission source. In this study, we used a simple permutation test to explore the spatial correlations between the ambient air and soil polychlorinated dibenzo-p-dioxin/dibenzofuran (PCDD/F) congener profile concentrations and their distances to a municipal solid waste incinerator (MSWI) in northern Taiwan. Kendall’s τ was applied for the correlation between the PCDD/F concentrations of a total of 9 airborne and 12 soil sampling sites and their distances to the incinerator. The results showed that the correlation was highly negatively associated for soil PCDD/Fs (τ = −0.515, p-value = 0.011) and was marginal for airborne PCDD/Fs (τ = −0.667, p-value = 0.055), conditional on the seasonal wind attributions. For the 17 individual congener profiles, most of the air and soil samples had a negative association, with 7 and 10 reaching statistical significance (p-value < 0.05) or borderline significance (p-value < 0.10), respectively. The permutation test provides a simple and robust statistical method for the identification of a potential pollution source from a limited sample size. Our analytical results show that dioxin emissions from the investigated MSWI had a definite environmental impact on the surrounding area.     

Point source identification using a simple permutation test: a case study of elevated PCDD/F levels in ambient air and soil and their relation to the distance to a local municipal solid waste incinerator

A point source for environmental pollution may be identified by the spatial association of samples collected in the vicinity of the emission source. In this study, we used a simple permutation test to explore the spatial correlations between the ambient air and soil polychlorinated dibenzo-p-dioxin/dibenzofuran (PCDD/F) congener profile concentrations and their distances to a municipal solid waste incinerator (MSWI) in northern Taiwan. Kendall’s τ was applied for the correlation between the PCDD/F concentrations of a total of 9 airborne and 12 soil sampling sites and their distances to the incinerator. The results showed that the correlation was highly negatively associated for soil PCDD/Fs (τ = −0.515, p-value = 0.011) and was marginal for airborne PCDD/Fs (τ = −0.667, p-value = 0.055), conditional on the seasonal wind attributions. For the 17 individual congener profiles, most of the air and soil samples had a negative association, with 7 and 10 reaching statistical significance (p-value < 0.05) or borderline significance (0.05 < p-value < 0.10), respectively. The permutation test provides a simple and robust statistical method for the identification of a potential pollution source from a limited sample size. Our analytical results show that dioxin emissions from the investigated MSWI had a definite environmental impact on the surrounding area.     

Love wave dispersion: A comparison of results for a semi-infinite medium with inhomogeneous layers and for its approximation by homogeneous layers

Love wave dispersion in various semi-infinite media consisting of inhomogeneous layers is discussed. The phase and group velocities are computed when shear wave velocity and density in each inhomogeneous layer are varying exponentially with depth. At the beginning one or two inhomogeneous layers over a homogeneous semi-infinite medium are considered. The dispersion results for these structures are compared with those for their approximations with homogeneous layers. Comparisons show that differences of phase and group velocities for the original models from those for their approximated models (i) increase with the increase of wave number and (ii) are larger for group velocity than for phase velocity. The difference is approximately proportional to the rate of change of parameters in the layers. Finally, dispersion curves are obtained for model IP3MC, which consists of many inhomogeneous and homogeneous layers over a homogeneous semi-infinite medium. The results are compared with the observed group velocity data across the Indian Peninsula.     

Aerial magnetic mapping with an unmanned aerial vehicle and a fluxgate magnetometer: a new method for rapid mapping and upscaling from the field to regional scale

Magnetic measurements with an unmanned aerial vehicle are ideal for filling the gap between ground and airborne magnetic surveying. However, to obtain accurate aeromagnetic data, the compensation of magnetic effects of the unmanned aerial vehicle is a challenge. Typically, scalar magnetometers are towed several metres under the unmanned aerial vehicle to minimize its magnetic field. In this study, a fluxgate three-component magnetometer is attached 42 cm in front of the unmanned aerial vehicle at the tip of a composite pipe. Using a scalar calibration, the sensor can be calibrated, and the permanent and induced magnetic fields of the unmanned aerial vehicle can be compensated. The contributions of the magnetic measurements at different altitudes to the unmanned aerial vehicle results were tested over an area of 1 km² in the Northern Vosges Mountains. The area is located in a hamlet surrounded by a forest where few geological outcrops are observed. Three magnetic surveys of the same area are obtained at different altitudes: 100, 30 and 1 m above the ground. The unmanned aerial vehicle magnetic data are compared with a helicopter aeromagnetic survey at 300 m above the ground and a ground magnetic survey using upward continuations of the maps to compare the results. The magnetic maps (300, 100, 30 and 1 m above the ground) show very different magnetic anomaly patterns (e.g. amplitude, shape, wavelength and orientation). The magnetic data at different altitudes improve the understanding of the geology from the local to more general scales.     

Geomorphic characteristics of the Minjiang drainage basin (eastern Tibetan Plateau) and its tectonic implications: New insights from a digital elevation model study

Abstract   The Minshan Mountain and adjacent region are the major continental escarpments along the eastern Tibetan Plateau. The Minjiang drainage basin is located within the plateau margin adjacent to the Sichuan Basin. Based on the analysis of the digital elevation model (DEM) acquired by the Shuttle Radar Topography Mission (SRTM), we know that the Minjiang drainage basin has distinct geomorphic characteristics. The regular increasing of local topographic relief from north to south is a result of the Quaternary sediment deposition within the plateau and the holistic uplift of the eastern margin of the Tibetan Plateau versus the Sichuan Basin. Results from DEM-determined Minjiang drainage sub-basins and channel profiles show that the tributaries on the opposite sides are asymmetric. Lower perimeter and area of drainage sub-basins, total channel length and bifurcation ratio within eastern flank along the Minjiang mainstream are the result of the Quaternary differential uplift of the Minshan Mountain region. Shorter stream lengths and lower bifurcation ratio might be the indications of the undergrowth and newborn features of these eastern streams, which are also representative for the eastern uplift of the Minshan Mountain.     

The phosphoric acid leak from the wreck of the MV Ece in the English Channel in 2006: Assessment with a ship of opportunity, an operational ecosystem model and historical data

This study evaluates the ship of opportunity (Ferrybox) concept for both sustained monitoring of UK shelf sea waters and numerical model validation. Release of phosphate from the wreck of a chemical tanker (MV Ece) in the western English Channel (49.73 degrees N, 3.25 degrees W) in March 2006 is used to demonstrate the importance of sustained observations in decision support systems and policy development. The Ferrybox system continuously collects sea surface (5m) data from a suite of autonomous electronic sensors installed on a passenger ferry operating year-round between Portsmouth (UK) and Bilbao (Spain). The detection of anomalously high concentrations of phosphate (1.54mmolm(-3), four times the usual level) and onset of phytoplankton growth close to the wreck site in March 2006 was placed in the context of multiple years of measurements (phosphate, nitrate, silicate and chlorophyll) collected from the Ferrybox system (2003-2006) and the long-term time series station E1 (50.03 degrees N, 4.65 degrees W, 1930-1987) in the English Channel. With regard to decision support, release of phosphate from the tanker is unlikely to pose a threat as phytoplankton growth at the end of winter is not unusual in this region and dissolved inorganic nitrogen rather than phosphate (DIN:DIP=10-18) is likely to ultimately limit algal growth in spring 2006. With regard to policy development, the Oslo and Paris (OSPAR) commissions recommendation of sampling every three years in "non-problem areas" is likely to provide statistically inadequate data, given the interannual and decadal variability identified in the Ferrybox and E1 data: the Ferrybox data show that oceanic winter nutrient concentrations varied by 35-50% between 2003/2004 and 2005/2006 due to deeper mixing of water off-shelf in early 2005/2006 and comparisons between the Ferrybox and E1 years show that the western English Channel is currently experiencing a low in phosphate concentrations similar to those in the 1960s. The importance of Ferrybox data in evaluating the reliability of predictive operational models needed in decision support is also demonstrated, by highlighting both strengths and weaknesses in a state-of-the-art ecosystem model designed for UK shelf waters.