Coupled atmosphere–ocean data assimilation experiments with a low-order climate model

A simple idealized atmosphere–ocean climate model and an ensemble Kalman filter are used to explore different coupled ensemble data assimilation strategies. The model is a low-dimensional analogue of the North Atlantic climate system, involving interactions between large-scale atmospheric circulation and ocean states driven by the variability of the Atlantic meridional overturning circulation (MOC). Initialization of the MOC is assessed in a range of experiments, from the simplest configuration consisting of forcing the ocean with a known atmosphere to performing fully coupled ensemble data assimilation. “Daily” assimilation (that is, at the temporal frequency of the atmospheric observations) is contrasted with less frequent assimilation of time-averaged observations. Performance is also evaluated under scenarios in which ocean observations are limited to the upper ocean or are non-existent. Results show that forcing the idealized ocean model with atmospheric analyses is inefficient at recovering the slowly evolving MOC. On the other hand, daily assimilation rapidly leads to accurate MOC analyses, provided a comprehensive set of oceanic observations is available for assimilation. In the absence of sufficient observations in the ocean, the assimilation of time-averaged atmospheric observations proves to be more effective for MOC initialization, including the case where only atmospheric observations are available.     

Assimilation of satellite altimeter data in a primitive-equation model of the Azores–Madeira region

The aim of this study is to implement satellite altimetric assimilation into a high-resolution primitive-equation ocean model and check the validity and sensitivity of the results. Beyond this paper, the remote objective is to get a dynamical tool capable of simulating the surface ocean processes linked to the air–sea interactions as well as to perform mesoscale ocean forecasting. For computational cost and practical reasons, this study takes place in a 1000 by 1000 sq km open domain of the Canary basin. The assimilation experiments are carried out with the combined TOPEX/POSEIDON and ERS-1 data sets between June 1993 and December 1993. The space–time domain overlaps with in situ data collected during the SEMAPHORE experiment and thus enables an objective validation of the results. A special boundary treatment is applied to the model by creating a surrounding recirculating area separated from the interior by a buffer zone. The altimetric assimilation is done by implementing a reduced-order optimal interpolation algorithm with a special vertical projection of the surface model/data misfits. We perform a first experiment with a vertical projection onto an isopycnal EOF representing the Azores Current vertical variability. An objective validation of the model's velocities with Lagrangian float data shows good results (the correlation is 0.715 at 150 dbar). The question of the sensitivity to the vertical projection is addressed by performing similar experiments using a method for lifting/lowering of the water column, and using an EOF in Z-coordinates. Some comparisons with in situ temperature data do not show any significant difference between the three projections, after five months of assimilation. However, in order to preserve the large-scale water characteristics, we felt that the isopycnal projection was a more physically consistent choice. Then, the complementary character of the two satellites is assessed with two additional experiments which use each altimeter data sets separately. There is an evidence of the benefit of combining the two data sets. Otherwise, an experiment assimilating long-wavelength bias-corrected CLS altimetric maps every 10 days exhibits the best correlation scores and emphasizes the importance of reducing the orbit error and biases in the altimetric data sets. The surface layers of the model are forced using realistic daily wind stress values computed from ECMWF analyses. Although we resolve small space and time scales, in our limited domain the wind stress does not significantly influence the quality of the results obtained with the altimetric assimilation. Finally, the relative effects of the data selection procedure and of the integration times (cycle lengths) is explored by performing data window experiments. A value of 10 days seems to be the most satisfactory cycle length.     

How dynamical models can learn from the data��an example with a simplified ENSO model

Most dynamical models of the natural system contain a number of empirical parameters which reflect our limited understanding of the simulated system or describe unresolved subgrid-scale processes. While the parameterizations basically introduce some uncertainty to the model results, they also hold the prospect of tuning the model. In general, a deterministic tuning is related to an inversion of the model which is often impossible or requires considerable computing effort for most climate models. Another way to adjust the model parameters to a specific observed process is stochastic fitting where a set of parameters and model output are taken as random variables. Here, we present a dynamical?Cstatistical approach with a simplified model of the El Ni?o?CSouthern Oscillation (ENSO) cycle whose parameters are adjusted to simulated and observed data by means of Bayesian statistics. As ENSO model, we employ the Schop?CSuarez delay oscillator model. Monte Carlo experiments highlight the large sensitivity of the model results to varied model parameters and initial values. The statistical adjustment is done by Bayesian model averaging of the Monte Carlo experiments. Applying the method to simulated data, the posterior ensemble mean is much closer to the reference data than the prior ensemble mean. The learning effect of the model is evident in the leading empirical orthogonal functions and statistically significant in the mean state. When the method is applied to the observed ENSO time series, the ENSO model in its classical setup is not able to account for the temporally varying periodicity of the observed ENSO phenomenon. An improved setup with continuous adjustment periods and extended parameter range is developed in order to allow the model to learn from the data gradually. The improved setup leads to promising results during the twentieth century and even a weak forecast skill over 6?months. Thus, the described method offers a promising tool for data assimilation in dynamical weather and climate models. However, the simplified ENSO model is barely appropriate for operational ENSO forecasts owing to its limited physical complexity.     

A weakly coupled data assimilation system of a coupled physical–biological model for the northeastern South China Sea

南海东北部沿岸海域在每年夏季都存在显著的物理过程(如上升流,珠江冲淡水)和活跃的生物过程,这些基本的物理场和生物场要素对沿岸海域的研究起关键作用。本文在一适应于南海东北部的物理生物耦合模式的基础上建立了一个弱约束的EnOI耦合同化系统。开展的SST同化试验结果表明,物理过程存在合理的响应,且近海上升流区的海洋生态变量对物理变量的同化也有响应。     

Regional climate model simulations of daily maximum and minimum near-surface temperatures across Europe compared with observed station data 1961–1990

Simulations of the present-day temperature climate in Europe obtained with the dynamic regional climate model HadRM3P from the Hadley Centre are evaluated. Observed daily temperature maxima (T _x) and minima (T_n) for the 1961–1990 period at 185 stations are compared with their nearest corresponding HadRM3P grid-box data. The model generally performs well over the UK and elsewhere between latitudes 50 and 55°N, with biases mostly within ±0.5 K. In other areas coherent regions with seasonal biases up to more than ±5 K exist. In some areas, biases in climatological averages are associated with even larger errors (up to more than ±15 K) in the upper/lower extreme temperature range. Both areas with systematically overestimated and underestimated intra-seasonal daily temperature variances exist, but overestimation dominates. Too hot summer T _x south of about 45°N are associated with drying soils in the model. This problem may occur further north in future integrations with a greenhouse-gas induced warming. Given the existence of errors in the simulations of the present-day climate, we recommend that results from future scenario integrations are treated with care.     

Study on the “winter persistence barrier” of Indian Ocean dipole events using observation data and CMIP5 model outputs

This study investigates the persistence barrier phenomenon associated with positive Indian Ocean dipole (IOD) events during the various phases of its development. The results derived from three observational datasets (the Simple Ocean Data Assimilation, International Comprehensive Ocean–Atmosphere Data Set, and Extended Reconstructed Sea Surface Temperature) indicate that significant winter persistence barriers (WPBs) occur during IOD events, both in its growing and decaying phases. The simulation skill of the 14 models within the Coupled Model Intercomparison Project Phase 5 with respect to persistence barriers was also evaluated and compared with observational data. The results show that although most models were able to simulate the WPB reasonably well during the growing phase, only five models could capture the appropriate WPB during the decaying phase. Further analysis demonstrates that the zonal equatorial gradient of climatological sea surface temperature (SST) and zonal sea surface winds at the equator in the Indian Ocean are very weak in winter, which indicates that the coupling between ocean and atmosphere is weakest in winter and encourages a rapid variation of IOD events and a swift reduction of persistence, favoring the occurrence of WPBs; furthermore, a deep climatological thermocline in winter implies that the subsurface water temperature cannot influence SST readily, and the memory of the subsurface temperature cannot help SST to recover from the loss of persistence during this period, leading to the occurrence of WPBs. In addition, an analysis of the climatological conditions in the outputs from the 14 models shows that those models that can (cannot) capture the winter climatological conditions frequently simulate the WPBs appropriately (poorly). This confirms that the occurrence of the WPB for IOD events may be closely related to particular winter climatological conditions, indicating that the WPB is an inherent property of IOD events.     

A comparison of global tropospheric teleconnections using observed satellite and general circulation model total ozone column data for 1979–91

 The total ozone column is well correlated with tropospheric fields such as the heights of the upper tropospheric geopotential surfaces and thus it can provide useful information on temporal variability in the troposphere. The global availability of long period satellite measurements of the total ozone column, taken by the TOMS instruments since 1978, provides a valuable and independent data set for use in studies of seasonal and interannual climate variability. In this study, the global low-frequency seasonal teleconnections in the observed TOMS data from 1979–91 have been investigated using seasonal teleconnectivity maps and empirical orthogonal function analysis. They have also been compared with the results from a simulation made with the atmospheric GCM at Météo-France, having prescribed observed sea surface temperatures for the same period. In the observed total ozone, strong ENSO-related wave number one longitudinal dipole patterns are seen in both the tropics and in the Southern Hemisphere extratropics. The model shows much weaker variability in total ozone yet appears to be able to capture similar teleconnection patterns in the tropics related to ENSO. In the SH extratropics, the model total ozone shows a strong wave number 3 response rather than the wave number one dipole seen in the observations. A wave number 3 response is also evident in the 200 hPa geopotential height simulated by the model and in the NCEP analysis, and is consistent with the response in a linearised barotropic model forced in the Indonesian region. The different responses in the modelled and observed total ozone, suggest that tropopause effect is not the major factor in the SH extratropics, and it is likely that horizontal ozone transport also plays a role in this region. Despite a generally poor simulation of the zonal mean total ozone, the model was able to capture the anomalous strengthening of the SH stationary waves during austral spring of 1988, related to an intense stratosphere sudden warming. Received: 21 October 1996 / Accepted: 11 September 1997     

The UHF wind profiler at Vienna airport – data quality control and comparisons to rawinsonde data

Summary A 1290MHz wind profiler (Radian Lap-3000), at present one of three operational wind profilers in Austria, is operated at Vienna airport. In spite of quality assurance procedures as consensus averaging included in the data evaluation process from profiler raw data, some spurious peaks of wind speed and unrealistic changes of the wind vector in time or height occur in the wind measurements. This is especially true for sampling intervals of only 5 minutes which are used to resolve the temporal evolution of summer thunderstorms and frontal passages. Averaging periods of only a few minutes are rather the lower limit apt for wind profiler observations and result in a low data availability of 28%, whereas about 55% of data (relative to the maximum height range according to the parameter setting) are available for 10 to 30 minutes profiles.Approaches to a posteriori quality control using checks for automatic error detection are proposed and tested on a one and a half year data-set: Flagging data when the three-dimensional wind divergence exceeds a predefined limit (0.5s^–1) is in most cases successful in combination with thresholds for wind speed (2 times the median of the daily data-set) or wind shear (0.2s^–1).The wind profiler data is compared to wind profiles from the next radiosonde station where soundings are launched 4 times a day at Hohe Warte, approx. 20km northwest, at the hill-side of the Viennese Woods. Deviations of about 1m s^–1 in wind speed are found between the observations of the two systems. Differences between the wind profiles within the boundary layer can be explained by local differences in the wind regime observed at the airport and the radiosounding – blocking effects of the Viennese Woods during south-easterly flow. Comparing the profiler data to radiosoundings on a monthly basis gives a tool to monitor the profiler performance.     

The KNMI HEXMAX stress data – a reanalysis

An earlier discussion of the wind stress measurements of the 1986 HEXMAX experiment off the Dutch coast left some uncertainty about the actual relationship between the stress and the wave field. In this paper we try to find a more definitive answer by looking at the consequences for gradients of wind speed, stress and phase speed of the waves if we assume the Charnock coefficient to be either a constant or to have an inverse wave age dependence. It turns out that both assumptions have a wind speed range where they produce coherent results, but that they break down when the wave length exceeds a value that is related to the water depth.     

Parameter identification for Kalman filtering of noisy and censored data: Is this a data quality control method?

Summary The quality control of rain amount data is usually based on a posteriori defined sets of empirical thresholds, e.g. when crosscheking neighbouring observations. While these procedures are well suited for very long time series, it is difficult to adopt the rules for time series of special observation periods with higher resolution in time. We propose a linear Kalman filter and present several methods, how to fit it onto the stochastic structure of the time series in the presence of many zeroes. These zeroes may indicate light rain or no rain and, therefore, are considered censored. They do not fit into the stochastic structure of the non-zero values and have to be treated separately. Fitting one out of four model parameters of the Kalman filter also determinates two other dependent ones. Only one model parameter has to be known from independent sources. The fitting algorithms are tested with artificial rain rate time series with known stochastic structure and several zero rain data points. Furthermore, applied to time series of observed hourly rain amounts for 4 consecutive winter months, the Kalman filter shows its sensitivity for faulty data points. The detection of conspicuous data by the method of Kalman filtering is discussed.With 3 FiguresProf. Dr. H.-D. Schilling died November 22, 1997.     

Seasonal climate hindcasts with Eta model nested in CPTEC coupled ocean–atmosphere general circulation model

This work evaluates the added value of the downscaling technique employed with the Eta model nested in the CPTEC atmospheric general circulation model and in the CPTEC coupled ocean?Catmosphere general circulation model (CGCM). The focus is on the austral summer season, December?CJanuary?CFebruary, with three members each year. Precipitation, latent heat flux, and shortwave radiation flux at the surface hindcast by the models are compared with observational data and model analyses. The global models generally overestimate the precipitation over South America and tropical Atlantic. The CGCM and the nested Eta (Eta + C) both produce a split in the ITCZ precipitation band. The Eta + C produces better precipitation pattern for the studied season. The Eta model reduces the excessive latent heat flux generated by these global models, in particular the Eta + C. Comparison against PIRATA buoys data shows that the Eta + C results in the smallest precipitation and shortwave radiation forecast errors. The Eta + C comparatively best results are though as a consequence of both: the regional model resolution/physics and smaller errors on the lateral boundary conditions provided by the CGCM.     

A data assimilation technique based on the π-algorithm

A practical implementation of the data assimilation algorithm based on the Kalman filter in its complete formulation is impossible due to high dimension of the associated equation sets and to nonlinearity of the predicted processes. The main direction in the implementation of the Kalman filter is an ensemble approach. Under the assumption of ergodicity of random forecast errors, an alternative algorithm with respect to the ensemble Kalman filter can be considered, in which probability averaging is replaced by time averaging. The proposes algorithm is based this assumption. The algorithm is easy to implement; however, its convergence, applicability to the data assimilation problems, and connection to the Kalman filter have not been studied. In the paper, applicability of the π-algorithm to data assimilation is considered on an example of a simple one-dimensional advection equation. Use of this simple equation allows comparing the classical Kalman filter algorithm with various practical approaches to its implementation.     

Does the model regional bias affect the projected regional climate change? An analysis of global model projections

An analysis is presented of the dependence of the regional temperature and precipitation change signal on systematic regional biases in global climate change projections. The CMIP3 multi-model ensemble is analyzed over 26 land regions and for the A1B greenhouse gas emission scenario. For temperature, the model regional bias has a negligible effect on the projected regional change. For precipitation, a significant correlation between change and bias is found in about 30% of the seasonal/regional cases analyzed, covering a wide range of different climate regimes. For these cases, a performance-based selection of models in producing climate change scenarios can affect the resulting change estimate, and it is noted that a minimum of four to five models is needed to obtain robust precipitation change estimates. In a number of cases, models with largely different precipitation biases can still produce changes of consistent sign. Overall, it is assessed that in the present generation of models the regional bias does not appear to be a dominant factor in determining the simulated regional change in the majority of cases.     

Short‐term forecasting with a multi‐level spectral primitive equation model part I ‐ model formulation

A global baroclinic primitive equation model using the spectral technique has been constructed for short‐ and medium‐range numerical weather prediction. The spectral technique, which is a special case of the Galerkin method, employs spherical harmonic basis functions in the evaluation of all horizontal derivatives. The use of a transform technique allows all the horizontal operations to be performed efficiently and allows physical processes to be evaluated in real space. The model employs a semi‐implicit algorithm for time integration and finite differencing in the vertical. Physical processes include orography, moist convection, large scale precipitation and boundary layer processes.     

Parametric model of the Earth’s radiation budget

Summary Proposition of the Earths radiation budget, based on properties of the absorption, reflection and scattering coefficients, is presented. In the model, the final fluxes of radiation are expressed directly by absorption properties of the atmospheres components. Consequently, total (i.e., with all feedback) greenhouse effect, forced by any component of the atmosphere, can be determined. Presented model is averaged and one dimensional, but it can be developed if it meets with acceptance.     

An observational‐data based evapotranspiration function for general circulation models

Abstract

In this paper Nappo’ s (1975) formulation of the moisture availability function was used to derive the ß function of the form employed in the evapotranspiration calculations of various GCMs (Carson, 1981). An inverse calculation using the planetary boundary‐layer parameterizations of the GLAS general circulation model was made to derive this function. For this purpose two ground temperatures, namely those of saturated and naturally dry ground, were prognostically carried in a 47‐day integration with the model. The form of the calculated ß function is different from those reviewed by Carson. An example of global evapotranspiration obtained with the derived ß function is shown. Two separate five‐day simulations, one using ß as derived here and the other using Nappo's (1975) M function as a ß function, are compared. Large differences in the calculated evapotranspiration occur in dry regions.     

Comment on “application of the Canadian regional climate model to the Laurentian Great Lakes region: Implementation of a lake model”

Abstract

A recently published slab model formulation of lake thermodynamics (Goyette et al., 2000), including an empirical factor to adjust the incoming heat flux so that the modelled lake surface temperature agrees with observed climatology, leads to a distinct lack of energy conservation. The empirical adjustment conceptually represents an exchange of heat between the mixed‐layer water (the slab that is explicitly simulated in the model) and deeper layers of water. It ensures a realistic temporal progression of temperature in the mixed layer, but the thermodynamic balance of the deeper water is not considered. When the deeper water is considered, it is found that the empirical adjustment accounts for the entire heat input to the deeper water, and on an annual mean basis, it is considerably unbalanced. This reveals a flaw in this model concept and, although not entirely invalidating the model, it needs to be included as a caveat in its use.     

Ice‐floe collisions interpreted from acceleration data during LIMEX ‘89

Abstract

Wave‐induced ice motions measured during the Labrador Ice Margin Experiment (LIMEX ‘89) were interpreted to determine the cause and the frequency of collisions between floes. The LIMEX acceleration data were acquired with an optimal resolution near the predominant wave frequency and did not contain information above 0.5 Hz. It was therefore possible to establish the frequency of collisions, but not the magnitude of the events. Events were defined by any contact between floes in a wave cycle, and the distribution of times between events indicates that floes are more likely to collide in adjacent wave cycles than if the events were independent. Periods of continuous and intermittent collisions were related to the wave characteristics, and the frequency of events increased with a decrease in air temperature and an increase in local wind speed. Contrary to expectations, there was not a positive relation between collision frequency and wave amplitude.     

The global climatology of an interannually varying air–sea flux data set

The air–sea fluxes of momentum, heat, freshwater and their components have been computed globally from 1948 at frequencies ranging from 6-hourly to monthly. All fluxes are computed over the 23 years from 1984 to 2006, but radiation prior to 1984 and precipitation before 1979 are given only as climatological mean annual cycles. The input data are based on NCEP reanalysis only for the near surface vector wind, temperature, specific humidity and density, and on a variety of satellite based radiation, sea surface temperature, sea-ice concentration and precipitation products. Some of these data are adjusted to agree in the mean with a variety of more reliable satellite and in situ measurements, that themselves are either too short a duration, or too regional in coverage. The major adjustments are a general increase in wind speed, decrease in humidity and reduction in tropical solar radiation. The climatological global mean air–sea heat and freshwater fluxes (1984–2006) then become 2 W/m² and ?0.1 mg/m² per second, respectively, down from 30 W/m² and 3.4 mg/m² per second for the unaltered data. However, decadal means vary from 7.3 W/m² (1977–1986) to ?0.3 W/m² (1997–2006). The spatial distributions of climatological fluxes display all the expected features. A comparison of zonally averaged wind stress components across ocean sub-basins reveals large differences between available products due both to winds and to the stress calculation. Regional comparisons of the heat and freshwater fluxes reveal an alarming range among alternatives; typically 40 W/m² and 10 mg/m² per second, respectively. The implied ocean heat transports are within the uncertainty of estimates from ocean observations in both the Atlantic and Indo-Pacific basins. They show about 2.4 PW of tropical heating, of which 80% is transported to the north, mostly in the Atlantic. There is similar good agreement in freshwater transport at many latitudes in both basins, but neither in the South Atlantic, nor at 35°N.