Seasonal and Extraseasonal Predictions of Summer Monsoon Precipitation by Gcms

ＳｅａｓｏｎａｌａｎｄＥｘｔｒａｓｅａｓｏｎａｌＰｒｅｄｉｃｔｉｏｎｓｏｆＳｕｍｍｅｒＭｏｎｓｏｏｎＰｒｅｃｉｐｉｔａｔｉｏｎｂｙＧｃｍｓ①ＺｅｎｇＱｉｎｇｃｕｎ（曾庆存），ＹｕａｎＣｈｏｎｇｇｕａｎｇ（袁重光），ＬｉＸｕ（李旭），ＺｈａｎｇＲｏｎｇ...     

Model identification for hydrological forecasting under uncertainty

Methods for the identification of models for hydrological forecasting have to consider the specific nature of these models and the uncertainties present in the modeling process. Current approaches fail to fully incorporate these two aspects. In this paper we review the nature of hydrological models and the consequences of this nature for the task of model identification. We then continue to discuss the history (“The need for more POWER‘’), the current state (“Learning from other fields”) and the future (“Towards a general framework”) of model identification. The discussion closes with a list of desirable features for an identification framework under uncertainty and open research questions in need of answers before such a framework can be implemented.     

平均环流的诊断与预报

本文采用实际的大气资料分析了一个正压平均环流距平模式各项的相对重要性.结果表明,单纯考虑初始场的惯性预报的准确率随所取平均时间的增加而迅速下降,并以长波和超长波段最为显著. 加上气候平均风场及地球球面效应等动力因子后的修正的惯性预报,其结果不如纯惯性预报好.而且,所取的平均时间越长,效果越差.加进散度订正因子后,对长波和超长波的预报起明显改进作用, 对平均环流的预报起重要作用的因子是未知的强迫作用项.根据分析,本文提出了提高平均环流预报准确率的可能途径.     

Solar Cycle Predictions based on Solar Activity at Different Solar Latitudes

Many methods of predictions of sunspot maximum number use data before or at the preceding sunspot minimum to correlate with the following sunspot maximum of the same cycle, which occurs a few years later. Kane and Trivedi (Solar Phys. 68, 135, 1980) found that correlations of R _z(max) (the maximum in the 12-month running means of sunspot number R _z) with R _z(min) (the minimum in the 12-month running means of sunspot number R _z) in the solar latitude belt 20° – 40°, particularly in the southern hemisphere, exceeded 0.6 and was still higher (0.86) for the narrower belt > 30° S. Recently, Javaraiah (Mon. Not. Roy. Astron. Soc. 377, L34, 2007) studied the relationship of sunspot areas at different solar latitudes and reported correlations 0.95 – 0.97 between minima and maxima of sunspot areas at low latitudes and sunspot maxima of the next cycle, and predictions could be made with an antecedence of more than 11 years. For the present study, we selected another parameter, namely, SGN, the sunspot group number (irrespective of their areas) and found that SGN(min) during a sunspot minimum year at latitudes > 30° S had a correlation +0.78±0.11 with the sunspot number R _z(max) of the same cycle. Also, the SGN during a sunspot minimum year in the latitude belt (10° – 30° N) had a correlation +0.87±0.07 with the sunspot number R _z(max) of the next cycle. We obtain an appropriate regression equation, from which our prediction for the coming cycle 24 is R _z(max )=129.7±16.3.     

Capture zone,travel time,and solute-transport predictions using inverse modeling and different geological models

Six regional-scale flow models are compared to gain insight into how different representations of hydraulic-conductivity distributions affect model calibration and predictions. Deterministic geological models were used to define hydraulic-conductivity distributions in two steady-state flow models that were calibrated to heads and baseflow estimates using inverse techniques. Optimized hydraulic-conductivity estimates from the two models were used to calculate layer and model mean hydraulic-conductivity values. Despite differences in the two geological models, inverse calibration produced mean hydraulic-conductivity values for the entire model domain that are quite similar. The layer and model mean hydraulic-conductivity values were used to generate four additional flow models and forward runs were performed. All of the models adequately simulate the observed heads and total baseflow. The six flow models were used to predict the steady-state impact of a proposed well field, and the flow solutions were used in simulating particle tracking and solute transport. Results of the predictive simulations show that, for this example, simple models of heterogeneity produce capture zones similar to more complex models, but with very different travel times and breakthroughs. Inverse modeling combined with different geological models can provide a measure of capture zone and breakthrough reliability.

---

Resumen Se compara seis modelos de flujo a escala regional para conocer cómo afecta a la calibración y a la predicción del modelo diversas representaciones de la distribución de la conductividad hidráulica. Se ha utilizado modelos geológicos deterministas para definir las distribuciones de la conductividad hidráulica en dos modelos de flujo permanente, calibrados mediante técnicas inversas con niveles piezométricos y estimaciones del flujo de base. Se ha adoptado estimaciones optimizadas de la conductividad hidráulica de los dos modelos para calcular las cotas de las capas y sus conductividades hidráulicas medias. A pesar de las diferencias entre ambos modelos geológicos, con la calibración inversa se obtiene valores similares de conductividad hidráulica en todo el dominio. Estos valores de las capas y de las conductividades hidráulicas medias han servido para generar cuatro modelos adicionales de flujo y realizar predicciones. Todos los modelos simulan de forma adecuada los niveles observados y los caudales de base. Los seis modelos han sido aplicados a la predicción del impacto estacionario de un campo de pozos, y las soluciones del flujo permiten simular el transporte de partículas y de solutos. Los resultados de estas predicciones muestran que, para este ejemplo, los modelos sencillos de la heterogeneidad dan lugar a zonas de captura similares a las generadas por modelos más complejos, pero aparecen grandes diferencias en los tiempos de tránsito y en las curvas de llegada. Una combinación de modelación inversa y de modelos geológicos diferentes puede proporcionar una medida de la fiabilidad de la zona de captura y de las curvas de llegada.

---

Résumé Six modèles d'écoulement à l'échelle régionale sont comparés afin d'avoir un aperçu de la manière dont les différentes représentations de la distribution de la conductivité hydraulique affectent la calibration et les prédictions de modèles. Des modèles géologiques déterministes ont été utilisés pour définir les distributions de la conductivité hydraulique dans deux modèles d'écoulement en régime permanent qui ont été calibrés avec des estimations des charges et des écoulements de base faites par des techniques inverses. Les estimations optimisées de la conductivité hydraulique de ces deux modèles ont servi à calculer les valeurs de conductivité hydraulique moyenne des couches et du modèle. Malgré des différences entre les deux modèles géologiques, la calibration inverse a donné des valeurs de conductivité hydraulique moyenne pour le domaine complet du modèle qui sont complètement semblables. Les valeurs de la conductivité moyenne des couches et du modèle ont été utilisées pour générer quatre modèles d'écoulement supplémentaires et des traitements ont été effectués. Tous les modèles simulent correctement les charges observées et l'écoulement de base total. Les six modèles ont servi à prédire l'impact en régime permanent d'un champ captant projeté et les solutions d'écoulement ont été utilisées dans une simulation par suivi de particules et de transport de soluté. Les résultats de simulations prédictives montrent que, pour cet exemple, de simples modèles d'hétérogénéité fournissent des zones de capture semblables aux modèles plus complexes, mais pour des temps de parcours et des restitutions très différents. Une modélisation inverse combinée à différents modèles géologiques peut assurer une mesure de la zone de capture et une fiabilité de la restitution.

---

     

Assessing agreement of measurements and predictions in geomorphology

Commonly in geomorphology measurements by different methods are compared to see how far they agree (i.e. are equal), as are predictions from models and corresponding observations. Such assessment usually employs scatter plots, correlation and possibly regression. More appropriate and more effective methods include plotting differences versus means and summary by concordance correlation and other measures of agreement. These methods, some new to geomorphology, are explained and discussed with a variety of examples using fluvial, hillslope, glacial and coastal data.     

How useful is the Waldmeier effect for prediction of a sunspot cycle?

Waldmeier effect [Waldmeier M., 1955. Ergebnisse und Probleme der Sonnenforschung. Second Ed., Leipzig, p. 154] states that the rise-time of a cycle depends upon a single parameter, namely the sunspot number Rz(max) at the maximum. Strong cycles have a steeper rise, while moderate cycles rise more slowly. In this paper, using the past data for sunspot cycles 1–23, these aspects are re-examined. It was noticed that the inverse relationship between Rz(max) and rise-time is discernable only when average patterns obtained by superposition of several cycles (separately for strong and weak cycles) are compared. In individual cycles, considerable deviations from the average patterns can occur (several tens of units of Rz and several months of rise-time). For a study of the relationship of Rz(max) with features in the early part of a cycle, the features chosen were Ro (i.e., Rz(min)) and Rz values Ra, Rb, and Rc, 12, 24 and 36 months, respectively, later than Ro (only 12-monthly running means were used). Ro had a moderate correlation (<0.6) with Rz(max), but Ra, Rb, Rc had better correlations. For hindsight predictions for cycles 18–23, the predictions for cycle 19 was grossly erroneous (observed value almost double of the predicted value). For other cycles, the errors were within 25%. For cycle 24, the Rz monthly values up to March 2008 give 12-month running means centered in June, July, August, September 2007 as 7.6, 6.5, 5.8, 6.1. Thus, though we cannot be absolutely sure yet that Rz(min) for cycle 24 has occurred, a tentative, provisional prediction using Rz(min) (i.e., Ro) as 5.8 is Rz(max)=113±19, i.e., in the range 94–132. This is an upper limit, as Ro value may reduce further in coming months, but most probably not very much. For Ro=5.0, the prediction would be Rz(max)= 109±17, while in the extreme hypothetical case of Ro=0.0, the prediction would be Rz(max)=79±14.     

Predictive vegetation mapping approach based on spectral data, DEM and generalized additive models

This study aims to provide a predictive vegetation mapping approach based on the spectral data, DEM and Generalized Additive Models (GAMs). GAMs were used as a prediction tool to describe the relationship between vegetation and environmental variables, as well as spectral variables. Based on the fitted GAMs model, probability map of species occurrence was generated and then vegetation type of each grid was defined according to the probability of species occurrence. Deviance analysis was employed to test the goodness of curve fitting and drop contribution calculation was used to evaluate the contribution of each predictor in the fitted GAMs models. Area under curve (AUC) of Receiver Operating Characteristic (ROC) curve was employed to assess the results maps of probability. The results showed that: 1) AUC values of the fitted GAMs models are very high which proves that integrating spectral data and environmental variables based on the GAMs is a feasible way to map the vegetation. 2) Prediction accuracy varies with plant community, and community with dense cover is better predicted than sparse plant community. 3) Both spectral variables and environmental variables play an important role in mapping the vegetation. However, the contribution of the same predictor in the GAMs models for different plant communities is different. 4) Insufficient resolution of spectral data, environmental data and confounding effects of land use and other variables which are not closely related to the environmental conditions are the major causes of imprecision.     

On accuracy of prediction of pre-failure deformations of granular soils

The problem of accuracy of predictions of pre-failure deformations of granular soils is considered. Firstly, the deterministic incremental equations describing qualitative character of such deformations are presented. These equations are based on extensive experimental data, obtained from the tri-axial compression tests. The parameters appearing in the incremental equations are assumed as random numbers, characterized by their mean values and standard deviations. The normal probability distribution is assumed. Then, such a model is applied to predict strains caused by some simple deterministic stress paths. The incremental equations are integrated for a large number of random coefficients, generated by a computer program. The final results are presented as either paths or points in the strain space, and respective density of probability distribution is calculated. The proposed method and results obtained show, that in mechanics of granular media we cannot expect “a good agreement” between predictions of theoretical models and experimental results, but only a kind of agreement measured by probability. The results presented should be important for both theoretical modellers and experimentalists.     

Seasonal and extraseasonal predictions of summer monsoon precipitation by gems(1)

A semi-operational real time short-term climate prediction system has been developed in the Center of Climate and Environment Prediction Research (CCEPRE), Institute of Atmospheric Physics / Chinese Academy of Sciences. The system consists of the following components: the AGCM and OGCM and their coupling, initial conditions and initialization, practical schemes of anomaly prediction, ensemble prediction and its standard deviation, correction of GCM output, and verification of prediction. The experiences of semi-operational real-time prediction by using this system for six years 1989-1994) and of hindcasting for 1980-1989 are reported. It is shown that in most cases large positive and negative anomalies of summer precipitation resulting in disastrous climate events such as severe flood or drought over East Asia can be well predicted for two seasons in advance, although the quantitatively statistical skill scores are only satisfactory due to the difficulty in correctly predicting the signs of small anomalies. Some methods for removing the systematic errors and introducing corrections to the GCM output arc suggested. The sensitivity of pre-diction to the initial conditions and the problem of ensemble prediction are also discussed in the paper