Estimates Of Potential Predictability Research Articles

Downslope Windstorm Forecasting: Easier with a Critical Level, but Still Challenging for High-Resolution Ensembles

Abstract Strong downslope windstorms can cause extensive property damage and extreme wildfire spread, so their accurate prediction is important. Although some early studies suggested high predictability for downslope windstorms, more recent analyses have found limited predictability for such winds. Nevertheless, there is a theoretical basis for expecting higher downslope wind predictability in cases with a mean-state critical level, and this is supported by one previous effort to forecast actual events. To more thoroughly investigate downslope windstorm predictability, we compare archived simulations from the NCAR ensemble, a 10-member mesoscale ensemble run at 3-km horizontal grid spacing over the entire contiguous United States, to observed events at 15 stations in the western United States susceptible to strong downslope winds. We assess predictability in three contexts: the average ensemble spread, which provides an estimate of potential predictability; a forecast evaluation based upon binary-decision criteria, which is representative of operational hazard warnings; and a probabilistic forecast evaluation using the continuous ranked probability score (CRPS), which is a measure of an ensemble’s ability to generate the proper probability distribution for the events under consideration. We do find better predictive skill for the mean-state critical-level regime in comparison to other downslope windstorm–generating mechanisms. Our downslope windstorm warning performance, calculated using binary-decision criteria from the bias-corrected ensemble forecasts, performed slightly worse for no-critical-level events, and slightly better for critical-level events, than National Weather Service high-wind warnings aggregated over all types of high-wind events throughout the United States and annually averaged for each year between 2008 and 2019.

Reemergence of Antarctic sea ice predictability and its link to deep ocean mixing in global climate models

Satellite observations show a small overall increase in Antarctic sea ice extent (SIE) over the period 1979–2015. However, this upward trend needs to be balanced against recent pronounced SIE fluctuations occurring there. In the space of 3 years, the SIE sank from its highest value ever reached in September 2014 to record low in February 2017. In this work, a set of six state-of-the-art global climate models is used to evaluate the potential predictability of the Antarctic sea ice at such timescales. This first multi-model study of Antarctic sea ice predictability reveals that the ice edge location can potentially be predicted up to 3 years in advance. However, the ice edge location predictability shows contrasted seasonal performances, with high predictability in winter and no predictability in summer. The reemergence of the predictability from one winter to next is provided by the ocean through its large thermal inertia. Sea surface heat anomalies are stored at depth at the end of the winter and influences the sea ice advance the following year as they resurface. The effectiveness of this mechanism across models is found to depend upon the depth of the mixed layer. One should be very cautious about these potential predictability estimates as there is evidence that the Antarctic sea ice predictability is promoted by deep Southern Ocean convection. We therefore suspect models with excessive convection to show higher sea ice potential predictability results due to an incorrect representation of the Southern Ocean.

An Alternative Estimate of Potential Predictability on the Madden–Julian Oscillation Phase Space Using S2S Models

This study proposes an alternative method to estimate the potential predictability without assuming the perfect model. A theoretical consideration relates a maximum possible value of the initial-value error to the covariance between analysis and bias-corrected ensemble-mean forecast. To test the method, the prediction limit of the Madden–Julian Oscillation (MJO) was evaluated, based on three pairs of reanalysis and forecast datasets provided by the European Centre for Medium-Range Weather Forecasting, the Japan Meteorological Agency and the National Centers for Environmental Prediction, participating in the subseasonal-to-seasonal prediction project. The results showed that the predictability was higher when MJO amplitude exceeded unity, consistent with the conventional method in which the error is evaluated as the ensemble-forecast spread. Moreover, the multimodel analysis was also conducted because the proposed method is readily applicable to the multimodel average of ensemble-mean forecasts. The phase dependency of the MJO’s potential predictability is also discussed.

Predictability and Forecast Skill in NMME

Abstract Forecast skill and potential predictability of 2-m temperature, precipitation rate, and sea surface temperature are assessed using 29 yr of hindcast data from models included in phase 1 of the North American Multimodel Ensemble (NMME) project. Forecast skill is examined using the anomaly correlation (AC); skill of the bias-corrected ensemble means (EMs) of the individual models and of the NMME 7-model EM are verified against the observed value. Forecast skill is also assessed using the root-mean-square error. The models’ representation of the size of forecast anomalies is also studied. Predictability was considered from two angles: homogeneous, where one model is verified against a single member from its own ensemble, and heterogeneous, where a model’s EM is compared to a single member from another model. This study provides insight both into the physical predictability of the three fields and into the NMME and its contributing models. Most of the models in the NMME have fairly realistic spread, as represented by the interannual variability. The NMME 7-model forecast skill, verified against observations, is equal to or higher than the individual models’ forecast ACs. Two-meter temperature (T2m) skill matches the highest single-model skill, while precipitation rate and sea surface temperature NMME EM skill is higher than for any single model. Homogeneous predictability is higher than reported skill in all fields, suggesting there may be room for some improvement in model prediction, although there are many regional and seasonal variations. The estimate of potential predictability is not overly sensitive to the choice of model. In general, models with higher homogeneous predictability show higher forecast skill.

Comparison of Seasonal Potential Predictability of Precipitation

Abstract Three methods for estimating potential seasonal predictability of precipitation from a single realization of daily data are assessed. The estimation methods include a first-order Markov chain model proposed by Katz (KZ), and an analysis of covariance (ANOCOVA) method and a bootstrap method proposed by the authors. The assessment is based on Monte Carlo experiments, ensemble atmospheric general circulation model (AGCM) simulations, and observation-based data. For AGCM time series, ANOCOVA produces the most accurate estimates of weather noise variance, despite the fact that it makes the most unrealistic assumptions about precipitation (in particular, it assumes precipitation is generated by a Gaussian autoregressive model). The KZ method significantly underestimates noise variance unless the autocorrelation of precipitation amounts on consecutive wet days is taken into account. Both AGCM and observation-based data reveal that the fraction of potentially predictable variance is greatest in the tropics, smallest in the extratropics, and undergoes a strong seasonal variation. The three methods give consistent estimates of potential predictability for 67% of the globe.

An examination of potential seasonal predictability in recent reanalyses

AbstractThis study examined potential seasonal predictability of precipitation and 2‐m temperature in the recent global reanalyses from 1979 to 2012. The reanalyses being investigated are R1, R2, ERA‐40, JRA25, ERA‐I, MERRA, CFSR and 20CR. When compared against Global Precipitation Climatology Project (GPCP) precipitation, ERA‐I and CFSR provide the best estimates of potential predictability, MERRA, R1 and R2 overestimate predictability and ERA‐40 and JRA25 are unrealistic in the tropics. Predictability estimates of 2‐m temperature from 20CR, JRA25, R1, MERRA, CFSR and R2 exhibit better agreement with the reanalysis ensemble mean than ERA‐40 and ERA‐I, which identify less predictability.

Potential seasonal predictability: Comparison between empirical and dynamical model estimates

Methods for estimating potential seasonal predictability from a single realization of daily data are validated against an ensemble of simulations from an atmospheric model driven by observed sea surface temperature, sea ice extent, and greenhouse gas concentration. The methods give surprisingly good estimates of potential predictability of seasonal precipitation despite the fact that the methods assume Gaussian distributions. For temperature, the methods systematically underestimate weather noise variance over land, often by a factor of 2 or more. This bias can be reduced by taking account of precipitation‐induced variability. These conclusions may be model dependent, and hence, confirmation in other models would be of interest. Nevertheless, for the state‐of‐the‐art atmospheric model used in this study, the results strongly support the validity of the single time series approach to estimating potential predictability and enhances our confidence in previous estimates of potential predictability based on observations alone.

Comparison of statistical estimates of potential seasonal predictability

Four methods for estimating potential seasonal predictability from a single time series are compared. The methods are: an analysis of variance procedure proposed by Shukla and Gutzler (SG), a spectral method proposed by Madden (MN), a bootstrap method proposed by the authors, and an analysis of covariance (ANOCOVA) method proposed by the authors. The time series used for comparison are taken from Monte Carlo simulations, an atmospheric general circulation model (AGCM), and reanalysis data. The comparison clearly reveals that SG systematically underestimates weather noise variance more strongly than the other methods and is therefore not a generally useful method. MN produces the least biased estimates of weather noise variance, but it tends to have a higher probability of identifying insignificant predictability than the other methods. Unfortunately, no simple, universally corrected statements can be made regarding the relative performances of MN, ANOCOVA, and bootstrap based on the AGCM output. Overall, the reanalysis‐based estimates of potential predictability of seasonal mean temperature derived from these methods is generally in accord with previous estimates, both in spatial structure and in magnitude. Omitting SG, the other three methods consistently identify about 80% of the globe as significantly predictable, and about 5% of the globe as insignificantly predictable. The remaining 15% of the globe, mostly over extratropical land, yields inconsistent assessments of potential predictability, indicating sensitivity to the assumptions underlying each of the methods. Interestingly, winter mean temperature over most of North America is found to be insignificantly predictable by all three methods.

Bootstrap estimated seasonal potential predictability of global temperature and precipitation

] Potentialpredictabilityofseasonalmeantemperatureandprecipitation is assessed using a moving blocks bootstrapmethod. The bootstrap method allows the potentialpredictability of seasonal means to be assessed even forautocorrelated, highly non‐Gaussian, intermittent data. Theresults reveal that the largest fraction of predictable variancefor both temperature and precipitation occur mainly overthe tropics where El Nino/Southern Oscillation dominatesthe interannual variability. Statistically significant potentialpredictability also is found in extratropics for temperature,particularly over most oceans and appreciable land areas.The potential predictability of temperature is generally smallerover land than over ocean and displays a significant annualcycle. Potential predictability of precipitation displays spottyand less continuous spatial patterns over extratropical regionsand also undergoes a significant annual cycle. The potentialpredictability estimates are generally consistent with previousstudies, but some inconsistency is also observed, such as thelack of significant potential predictability for temperature overNorthAmericanwinter.

Prediction of the Madden–Julian oscillation with the POAMA dynamical prediction system

Predictions of the Madden–Julian oscillation (MJO) are assessed using a 10-member ensemble of hindcasts from POAMA, the Australian Bureau of Meteorology coupled ocean–atmosphere seasonal prediction system. The ensemble of hindcasts was initialised from observed atmosphere and ocean initial conditions on the first of each month during 1980–2006. The MJO is diagnosed using the Wheeler-Hendon Real-time Multivariate MJO (RMM) index, which involves projection of daily data onto the leading pair of eigenmodes from an analysis of zonal winds at 200 and 850 hPa and outgoing longwave radiation (OLR) averaged about the equator. Forecasts of the two component (RMM1 and RMM2) index are quantitatively compared with observed behaviour derived from NCEP reanalyses and satellite OLR using the bivariate correlation skill, root-mean-square error (RMSE), and measures of the MJO amplitude and phase error. Comparison is also made with a simple vector autoregressive (VAR) prediction model of RMM as a benchmark. Using the full hindcast set, we find that the MJO can be predicted with the POAMA ensemble out to about 21 days as measured by the bivariate correlation exceeding 0.5 and the bivariate RMSE remaining below ~1.4 (which is the value for a climatological forecast). The VAR model, by comparison, drops to a correlation of 0.5 by about 12 days. The prediction limit from POAMA increases by less than 2 days for times when the MJO has large initial amplitude, and has little sensitivity to the initial phase of the MJO. The VAR model, on the other hand, shows a somewhat larger increase in skill for times of strong MJO variability and has greater sensitivity to initial phase, with lower skill for times when MJO convection is developing in the Indian Ocean. The sensitivity to season is, however, greater for POAMA, with maximum skill occurring in the December–January–February season and minimum skill in June–July–August. Examination of the MJO amplitudes shows that individual POAMA members have slightly above observed amplitude after a spin-up of about 10 days, whereas examination of the MJO phase error reveals that the model has a consistent tendency to propagate the MJO slightly slower than observed. Finally, an estimate of potential predictability of the MJO in POAMA hindcasts suggests that actual MJO prediction skill may be further improved through continued development of the dynamical prediction system.

Decadal potential predictability of twenty-first century climate

Decadal prediction of the coupled climate system is potentially possible given enough information and knowledge. Predictability will reside in both externally forced and in long timescale internally generated variability. The “potential predictability” investigated here is characterized by the fraction of the total variability accounted for by these two components in the presence of short-timescale unpredictable “noise” variability. Potential predictability is not a classical measure of predictability nor a measure of forecast skill but it does identify regions where long timescale variability is an appreciable fraction of the total and hence where prediction on these scale may be possible. A multi-model estimate of the potential predictability variance fraction (ppvf) as it evolves through the first part of the twenty-first century is obtained using simulation data from the CMIP3 archive. Two estimates of potential predictability are used which depend on the treatment of the forced component. The multi-decadal estimate considers the magnitude of the forced component as the change from the beginning of the century and so becomes largely a measure of climate change as the century progresses. The next-decade estimate considers the change in the forced component from the past decade and so is more pertinent to an actual forecast for the next decade. Long timescale internally generated variability provides additional potential predictability beyond that of the forced component. The ppvf may be expressed in terms of a signal-to-noise ratio and takes on values between 0 and 1. The largest values of the ppvf for temperature are found over tropical and mid-latitude oceans, with the exception of the equatorial Pacific, and some but not all tropical land areas. Overall the potential predictability for temperature generally declines with latitude and is relatively low over mid- to high-latitude land. Potential predictability for precipitation is generally low and due almost entirely to the forced component and then mainly at high latitudes. To the extent that the multi-model ppvf reflects both the behaviour of the actual climate system and the possibility of decadal prediction, the results give some indication as to where and to what extent decadal forecasts might be possible.

Representation and prediction of the Indian Ocean dipole in the POAMA seasonal forecast model

AbstractPredictive skill of the Indian Ocean Dipole (IOD) using the Bureau of Meteorology's POAMA seasonal forecast model is assessed in light of the model's ability to represent the coupled nature of the IOD. The analysis is based on a ten‐member ensemble of nine‐month forecasts that are initialized from observed ocean and atmosphere states once per month for the period 1982–2006. Some aspects of the IOD are well represented by the POAMA forecast model, including the seasonal phase locking of the development of the IOD but the IOD's horizontal structure is distorted and its amplitude weakens with increasing forecast lead time, while at the same time the IOD becomes spuriously overdependent on ENSO. These deficiencies in the representation of the IOD are attributed to model error (including horizontal resolution and development of biases in the mean state), which affects depiction of important coupled dynamics for the IOD as well as the role of remote forcing of the IOD by ENSO (which itself exhibits distortion of its spatial structure attributed to biases in the simulated mean state).In the light of these limitations of representing the IOD in the POAMA forecast model, skilful prediction of the peak phase of the IOD in September–November is limited to about a four‐month lead time, with a rapid drop in skill for forecasts initialized before June. Forecast skill for the IOD appears to be limited mainly by ability to predict development of SST anomalies in its eastern pole. Potential predictability estimates, assuming a perfect model, indicate great scope for improved prediction, although the IOD is much less predictable (potentially and in practice) than ENSO. The deficient depiction of the IOD at longer lead times means that estimates of potential predictability assuming a perfect model may be unreliable. Copyright © 2009 Royal Meteorological Society

The Impact of Air–Sea Interactions on the Predictability of the Tropical Intraseasonal Oscillation

Abstract This study investigates whether air–sea interactions contribute to differences in the predictability of the boreal winter tropical intraseasonal oscillation (TISO) using the NCEP operational climate model. A series of coupled and uncoupled, “perfect” model predictability experiments are performed for 10 strong model intraseasonal events. The uncoupled experiments are forced by prescribed SST containing different types of variability. These experiments are specifically designed to be directly comparable to actual forecasts. Predictability estimates are calculated using three metrics, including one that does not require the use of time filtering. The estimates are compared between these experiments to determine the impact of coupled air–sea interactions on the predictability of the tropical intraseasonal oscillation and the sensitivity of the potential predictability estimates to the different SST forcings. Results from all three metrics are surprisingly similar. They indicate that predictability estimates are longest for precipitation and outgoing longwave radiation (OLR) when the ensemble mean from the coupled model is used. Most importantly, the experiments that contain intraseasonally varying SST consistently predict the control events better than those that do not for precipitation, OLR, 200-hPa zonal wind, and 850-hPa zonal wind after the first 10 days. The uncoupled model is able to predict the TISO with similar skill to that of the coupled model, provided that an SST forecast that includes these intraseasonal variations is used to force the model. This indicates that the intraseasonally varying SSTs are a key factor for increased predictability and presumably better prediction of the TISO.

Multi‐model decadal potential predictability of precipitation and temperature

A first multi‐model estimate of the long timescale potential predictability of precipitation is obtained based on over 8000 years of data from the control simulations of 21 state‐of‐the‐art coupled climate models. The analysis also updates earlier estimates of the potential predictability of temperature to provide a consistent estimate for these basic climate parameters. Long timescale potential predictability is found mainly over the oceans at middle to high latitudes and predominantly where the surface is connected to the deeper ocean. Precipitation's modest potential predictability resembles an attenuated version of that for temperature on these timescales. Over land, predictability is largely absent for precipitation and comparatively weak for temperature where it is found over the northern and western parts of northern hemisphere land masses bordering the oceans. Regions exhibiting potential predictability indicate where we may hope to find predictive skill at long timescales and also points indirectly to the processes involved.

Simulation of South-Asian Summer Monsoon in a GCM

Major characteristics of Indian summer monsoon climate are analyzed using simulations from the upgraded version of Florida State University Global Spectral Model (FSUGSM). The Indian monsoon has been studied in terms of mean precipitation and low-level and upper-level circulation patterns and compared with observations. In addition, the model's fidelity in simulating observed monsoon intraseasonal variability, interannual variability and teleconnection patterns is examined. The model is successful in simulating the major rainbelts over the Indian monsoon region. However, the model exhibits bias in simulating the precipitation bands over the South China Sea and the West Pacific region. Seasonal mean circulation patterns of low-level and upper-level winds are consistent with the model's precipitation pattern. Basic features like onset and peak phase of monsoon are realistically simulated. However, model simulation indicates an early withdrawal of monsoon. Northward propagation of rainbelts over the Indian continent is simulated fairly well, but the propagation is weak over the ocean. The model simulates the meridional dipole structure associated with the monsoon intraseasonal variability realistically. The model is unable to capture the observed interannual variability of monsoon and its teleconnection patterns. Estimate of potential predictability of the model reveals the dominating influence of internal variability over the Indian monsoon region.

Medium-Range, Monthly, and Seasonal Prediction for Europe and the Use of Forecast Information

AbstractOperational probabilistic (ensemble) forecasts made at ECMWF during the European summer heat wave of 2003 indicate significant skill on medium (3–10 day) and monthly (10–30 day) time scales. A more general “unified” analysis of many medium-range, monthly, and seasonal forecasts confirms a high degree of probabilistic forecast skill for European temperatures over the first month. The unified analysis also identifies seasonal predictability for Europe, which is not yet realized in seasonal forecasts. Interestingly, the initial atmospheric state appears to be important even for month 2 of a coupled forecast.Seasonal coupled model forecasts capture the general level of observed European deterministic predictability associated with the persistence of anomalies. A review is made of the possibilities to improve seasonal forecasts. This includes multimodel and probabilistic techniques and the potential for “windows of opportunity” where better representation of the effects of boundary conditions (e.g., sea surface temperature and soil moisture) may improve forecasts. “Perfect coupled model” potential predictability estimates are sensitive to the coupled model used and so it is not yet possible to estimate ultimate levels of seasonal predictability.The impact of forecast information on different users with different mitigation strategies (i.e., ways of coping with a weather or climate event) is investigated. The importance of using forecast information to reduce volatility as well as reducing the expected expense is highlighted. The possibility that weather forecasts can affect the cost of mitigating actions is considered. The simplified analysis leads to different conclusions about the usefulness of forecasts that could guide decisions about the development of “end-to-end” (forecast-to-user decision) systems.

Seasonal Predictability and Spatial Coherence of Rainfall Characteristics in the Tropical Setting of Senegal

Abstract This study examines space–time characteristics of seasonal rainfall predictability in a tropical region by analyzing observed data and model simulations over Senegal. Predictability is analyzed in terms of the spatial coherence of observed interannual variability at the station scale, and within-ensemble coherence of general circulation model (GCM) simulations with observed sea surface temperatures (SSTs) prescribed. Seasonal mean rainfall anomalies are decomposed in terms of daily rainfall frequency and daily mean intensity. The observed spatial coherence is computed from a 13-station network of daily rainfall during the July–September season 1961–98 in terms of (i) interannual variability of a standardized anomaly index (i.e., the average of the normalized anomalies of each station), (ii) the external variance (i.e., the fraction of common variance among stations), and (iii) the number of spatiotemporal degrees of freedom. Spatial coherence of interannual anomalies across stations is found to be much stronger for seasonal rainfall amount and daily occurrence frequency, compared with daily mean intensity of rainfall. Combinatorial analysis of the station observations suggests that, for occurrence and seasonal amount, the empirical number of spatial degrees of freedom is largely insensitive to the number of stations considered, and is between 3 and 4 for Senegal. For daily mean intensity, by contrast, each station is found to convey almost independent information, and the number of degrees of freedom would be expected to increase for a denser network of stations. The GCM estimates of potential predictability and skill associated with the SST forcing are found to be remarkably consistent with those inferred from the observed spatial coherence: there is a moderate-to-strong skill at reproducing the interannual variations of seasonal amounts and rainfall occurrence, whereas the skill is weak for the mean intensity of rainfall. Over Senegal during July–September, it is concluded that (i) regional-scale seasonal amount and rainfall occurrence frequency are predictable from SSTs, (ii) daily mean intensity of rainfall is spatially incoherent and largely unpredictable at the regional scale, and (iii) point-score estimates of seasonal rainfall predictability and skill are subject to large sampling variability.

Potential predictability of the Asian summer monsoon on monthly and seasonal time scales

The potential predictability of the monthly and seasonal means during the Northern Hemisphere summer and winter is studied by estimating the signal-to-noise ratio. Based on 33 years of daily low-level wind observations and 24 years of satellite observations of outgoing long wave radiation, the predictability of the Asian summer monsoon region is contrasted with that over other tropical regions. A method of separating the contributions from slowly varying boundary forcing and internal dynamics (e.g., intraseasonal oscillations) that determine the predictability of the monthly mean tropical climate is proposed. We show that the Indian monsoon climate is only marginally predictable in monthly time scales as the contribution of the boundary forcing in this region is relatively low and that of the internal dynamics is relatively large. It is shown that excluding the Indian monsoon region, the predictable region is larger and predictability is higher in the tropics during northern summer. Even though the boundary forced variance is large during northern winter, the predictable region is smaller as the internal variance is larger and covers a larger region during northern winter (due to stronger intraseasonal activity). Consistent with the estimates of predictability of monthly means, estimates of potential predictability on seasonal time scales also indicate that predictability of seasonal mean Indian monsoon is limited.

How far ahead could we predict El Niño?

[1] The upper limit of the predictability of the El Nino Southern Oscillation (ENSO) is examined using ensembles of simulations of a coupled ocean-atmosphere global circulation model which has a relatively realistic ENSO cycle. By making small perturbations to the initial conditions and measuring the rate of divergence of nearby trajectories, it is found that given a perfect model and near perfect initial conditions, ENSO could be usefully predicted, on average, up to 8 months in advance (where “useful” is defined for a forecast with an anomaly correlation coefficient of greater than 0.6). This is at the low end of potential predictability estimates obtained using intermediate models. Some ensemble experiments do show potential predictability beyond 12 months, but in others small “errors” in the initial conditions can saturate in less than 6 months. The physical mechanisms which underlie the high and low predictable states are briefly examined with a view to predicting the reliability of ENSO forecasts.