Combined Empirical Orthogonal Function Research Articles

Role of the Australian High in Seasonal Phase Locking of the Indian Ocean Dipole

AbstractThis paper analyzes the effect of the Australian High (AH) on the seasonal phase locking of Indian Ocean Dipole (IOD) events. The anomalous strong AH associated with the positive phase of the Antarctic Oscillation can cause significant easterly wind anomalies and northward cross‐equatorial flow over the western Maritime Continent (MC) by strengthening the Australian winter monsoon during May–August. The AH‐associated easterly anomalies and northward cross‐equatorial flow can create thermodynamic air‐sea feedback and contribute to a significant cooling anomaly in the western MC and the tropical eastern Indian Ocean. Without considering the effect of ENSO, these processes contribute to the occurrence of positive IOD events, which begin in early summer, peak in late summer, and decay rapidly thereafter. The effect of ENSO can extend the peak period of IOD into the boreal autumn of that year. An anomalous weak AH corresponds to the occurrence and seasonal phase locking of negative IOD events. Through combined empirical orthogonal function analysis, we find that the effect of the AH can well explain the seasonal phase locking of 34 IOD events (40 in total), which provides an important theoretical basis for the prediction of IOD events.

Intercomparisons of methods for extracting the internal climate variability from the observed records over the Indo‐Pacific sector

AbstractA careful isolation of the externally forced component (EFC) and the internal climate variability (ICV) embedded in the observed records as well as the climate simulations is critical to investigate an actual response to the external radiative forcing and/or background dynamics in the ICV. Employing three different methods, we evaluate the EFCs contained in the observed sea surface temperature (SST) and sea level pressure (SLP) fields in the Indo‐Pacific region. After removing these EFCs, we obtain the respective ICVs as the remaining anomalies. The remaining SST and SLP anomalies are then evaluated on decadal time scales in a combined empirical orthogonal function (EOF) analysis of different spatial portions: the tropical Pacific, the Indian Ocean and the whole Indo‐Pacific region. After making statistical intercomparisons of the spatial patterns and associated time series of the EOF analyses, we found that the EFCs of the individual grid point values (GPVs) were appropriately estimated by regressing onto the multi‐model ensemble global mean surface temperature (GMSTMME) and were less well approximated by the conventional linear trend and the multi‐model ensemble mean of the simulated GPVs. The regressed SST anomalies of the individual historical simulations onto the GMSTMME were much larger than the observed anomalies, illustrating that the ICV‐to‐EFC variance ratio is a performance‐improvement indicator of climate models.

The 1994 Positive Indian Ocean Dipole Event as Investigated by the Transfer Routes of Oceanic Wave Energy

Abstract The present study investigates the interannual variability of the tropical Indian Ocean (IO) based on the transfer routes of wave energy in a set of 61-yr hindcast experiments using a linear ocean model. To understand the basic feature of the IO dipole mode, this paper focuses on the 1994 pure positive event. Two sets of westward transfer episodes in the energy flux associated with Rossby waves (RWs) are identified along the equator during 1994. One set represents the same phase speed as the linear theory of equatorial RWs, while the other set is slightly slower than the theoretical phase speed. The first set originates from the reflection of equatorial Kelvin waves at the eastern boundary of the IO. On the other hand, the second set is found to be associated with off-equatorial RWs generated by southeasterly winds in the southeastern IO, which may account for the appearance of the slower group velocity. A combined empirical orthogonal function (EOF) analysis of energy-flux streamfunction and potential reveals the intense westward signals of energy flux are attributed to off-equatorial RWs associated with predominant wind input in the southeastern IO corresponding to the positive IO dipole event. Significance Statement The present study gains a new insight into the mechanism of the Indian Ocean dipole events using a new diagnostic scheme for wave energy based on 61-yr hindcast experiments. The results have shown the existence of two sets of westward transfer of wave energy at the equator during 1994. One set of westward signals shows the same group velocity with theoretical equatorial Rossby waves that appear reasonably along the equator. The other set of westward signals at the equator represents a slightly slower group velocity than the theoretical equatorial Rossby waves, which is associated with abnormally extended southeasterly winds during the Indian Ocean dipole event.

The Intraseasonal Variations of the Leading Mode of Summer Precipitation Anomalies in Meiyu Area of East Asia

The intraseasonal variations of summer precipitation anomalies in the Meiyu area of East Asia are analyzed by applying a combined empirical orthogonal function (CEOF) of the latest meteorological reanalysis data ERA5 of European Center for Medium-Range Weather Forecasts for the period from 1991 to 2020, and the circulation structures and sources of variability of CEOF are also investigated. The first mode of the intraseasonal variations shows an in-phase pattern over the Meiyu area in June, July, and August, accounting for 22.2% of the total variance in the intraseasonal variations of summer precipitation anomalies. The positive (negative) CEOF1 is accompanied by the negative (positive) East Asia/Pacific pattern, including strong westerly wind anomalies in the upper troposphere and southwest monsoon in the lower troposphere, and the Western Pacific Subtropical High extending westward and its ridge line slightly south. The positive CEOF1 is preceded by decay of El Niño episodes, including the abnormal warm sea surface temperature anomalies (SSTAs) in the equatorial Central-Eastern Pacific in spring and warm SSTAs in the equatorial Indian Ocean in summer. The second mode shows an opposite precipitation anomaly in June and July, and the distribution in August is not significant. The corresponding geopotential height circulation of positive CEOF2 shows the large negative anomaly in the region north of 40° N and a positive anomaly over Japan in June, whereas the pattern reverses in July. At the same time, there is a radical reversion from abnormal eastly to westly wind in the upper troposphere. The structure of CEOF2 is somewhat induced by local SSTAs over the Northern Indian Ocean and South China Sea.

Assessment of the Ability of CMIP6 GCMS to Simulate the Boreal Summer Intraseasonal Oscillation Over Southeast Asia

The boreal summer intraseasonal oscillation (BSISO) plays an important role in the intraseasonal variability of a wide range of weather and climate phenomena across the region modulated by the Asian summer monsoon system. This study evaluates the strengths and weaknesses of 19 Coupled Model Intercomparison Project Phase 6 (CMIP6) models to reproduce the basic characteristics of BSISO. The models' rainfall and largescale climates are evaluated against GPCP and ERA5 reanalysis datasets. All models exhibit intraseasonal variance of 30–60-day bandpass-filtered rainfall and convection anomalies but with diverse magnitude when compared with observations. The CMIP6 models capture the structure of the eastward/northward propagating BSISO at wavenumbers 1 and 2 but struggle with the intensity and location of the convection signal. Nevertheless, the models show a good ability to simulate the power spectrum and coherence squared of the principal components of the combined empirical orthogonal function (CEOF) and can capture the distinction between the CEOF modes and red noise. Also, the result shows that some CMIP6 models can capture the coherent intraseasonal propagating features of the BSISO as indicated by the Hovmöller diagram. The contribution of latent static energy to the relationship between the moist static energy and intraseasonal rainfall over Southeast Asia is also simulated by the selected models, albeit the signals are weak. Taking together, some of the CMIP6 models can represent the summertime climate and intraseasonal variability over the study region, and can also simulate the propagating features of BSISO, but biases still exist.

A new perspective on ENSO-Indian summer monsoon rainfall relationship in a warming environment

A robust estimate of the ENSO-Indian summer monsoon rainfall (ISMR) relationship in a changing climate is critically important for projection of the ISMR predictability. In this study, we show that the increasing nonlinear trend of ISMR and global sea surface temperatures (SST) associated with the trend of the global warming mode leads to a ‘long-term’ component to the ENSO-ISMR relationship together with a ‘short-range’ component associated with the ENSO mode. Using historical and projected simulations by some state-of-the-art CMIP5 models, we separate the global warming mode from other oscillatory modes including the ENSO mode through a combined empirical orthogonal function (CEOF) analysis, supported by a signal-to-noise maximizing EOF analysis. We find that the ‘short-term’ ENSO-ISMR relationship, represented by significant negative correlation between JJAS NINO3.4 SST and ISMR does not change appreciably with increase in green-house gas (GHG) forcing while the ‘long-term’ relationship changes from weak to significantly positive. We demonstrate that in a warming environment, the global warming mode must be taken into account together with the ‘natural’ ENSO mode in order to predict ISMR with confidence and emphasize that predictability of ISMR at any given time must be reassessed taking into account contributions from both short and long-term ENSO-ISMR relationships.

Multivariate analysis on Indonesian forest fire using combined empirical orthogonal function and covariance matrices

Forest and land fires are annual events that occur very frequently during the dry season in several regions in Indonesia. One of the main triggers is low rainfall and extreme anomalies in those regions. The objective of the research was to find the correlation between rainfall and Forest fire based on Tropical Rainfall Measuring Mission (TRMM) rainfall data and Global Fire Emission Database (GFED) forest fire data. Both of these data are big data, this research utilized singular value decomposition to generate information on the correlation of fire patterns and rainfall spatially and temporally. Empirical Orthogonal Function (EOF) based on Singular value decomposition (SVD) was used to reduce large data, without removing main information from the data. SVD approach in this research was conducted using the covariance matrix and combined EOF. EOF mode contribution is able to explain the joint variability of the data set to the phenomenon of fire and its relation to rainfall in Indonesia. Spatial patterns in this analysis show that Kalimantan, Sumatra, and Papua emerged as endemic areas for fires in 1998, 2002, 2006, 2009 and 2015. The year of 2004 and 2014 was also recorded as years of severe forest and land fires but was relatively not as severe as mentioned earlier. The temporal pattern shows that the characteristics of Forest fire in Indonesia are dominant in the final months of the year, from June to October.

An Ionosphere Specification Technique Based on Data Ingestion Algorithm and Empirical Orthogonal Function Analysis Method

AbstractA data ingestion method in reproducing ionospheric electron density and total electron content (TEC) was developed to incorporate TEC products from the Madrigal Database into the NeQuick 2 model. The method is based on retrieving an appropriate global distribution of effective ionization parameter (Az) to drive the NeQuick 2 model, which can be implemented through minimizing the difference between the measured and modeled TEC at each grid in the local time‐modified dip latitude coordinates. The performance of this Madrigal TEC‐driven‐NeQuick 2 result is validated through the comparison with various International Global Navigation Satellite Systems Services global ionospheric maps and ionosonde data. The validation results show that a general accuracy improvement of 30–50% can be achieved after data ingestion. In addition, the empirical orthogonal function (EOF) analysis technique is used to construct a parameterized time‐varying global Az model. The quick convergence of EOF decomposition makes it possible to use the first six EOF series to represent over 90% of the total variances. The intrinsic diurnal variation and spatial distribution in the original data set can be well reflected by the constructed EOF base functions. The associated EOF coefficients can be expressed as a set of linear functions of F10.7 and Ap indices, combined with a series of trigonometric functions with annual/seasonal variation components. The NeQuick TEC driven by EOF‐modeled Az shows 10–15% improvement in accuracy over the standard ionosphere correction algorithm in the Galileo navigation system. These preliminary results demonstrate the effectiveness of the combined data ingestion and EOF modeling technique in improving the specifications of ionospheric density variations.

Long-Lead Seasonal Prediction of China Summer Rainfall Using an EOF–PLS Regression-Based Methodology*,+

Abstract Considerable year-to-year variability of summer rainfall exposes China to threats of frequent droughts and floods. Objective prediction of the summer rainfall anomaly pattern turns out to be very challenging. As shown in the present study, the contemporary state-of-the-art dynamical models’ 1-month-lead prediction of China summer rainfall (CSR) anomalies has insignificant skills. Thus, there is an urgent need to explore other ways to improve CSR prediction. The present study proposes a combined empirical orthogonal function (EOF)–partial least squares (PLS) regression method to offer a potential long-lead objective prediction of spatial distribution of CSR anomalies. The essence of the methodology is to use PLS regression to predict the principal component (PC) of the first five leading EOF modes of CSR. The preceding December–January mean surface temperature field [ST; i.e., SST over ocean and 2-m air temperature (T2m) over land] is selected as the predictor field for all five PCs because SST and snow cover, which is reflected by 2-m air temperature, are the most important factors that affect CSR and because the correlation between each mode and ST during winter is higher than in spring. The 4-month-lead forecast models are established by using the data from 1979 to 2004. A 9-yr independent forward-rolling prediction is made for the latest 9 yr (2005–13) as a strict forecast validation. The pattern correlation coefficient skill (0.32) between the observed and the 4-month-lead predicted patterns during the independent forecast period of 2005–13 is significantly higher than the dynamic models’ 1-month-lead hindcast skill (0.04), which indicates that the EOF–PLS regression is a useful tool for improving the current seasonal rainfall prediction. Issues related to the EOF–PLS method are also discussed.

Intraseasonal variability of the Atlantic Intertropical Convergence Zone during austral summer and winter

The Atlantic Intertropical Convergence Zone (A-ITCZ) exhibits variations on several time-scales and plays a crucial role in precipitation regimes of northern South America and western Africa. Here we investigate the variability of the A-ITCZ on intraseasonal time-scales during austral summer (November–March) and winter (May–September) based on a multivariate index that describes the main atmospheric features of the A-ITCZ and retains its variability on interannual, semiannual, and intraseasonal time-scales. This index is the time coefficient of the first combined empirical orthogonal function mode of anomalies (annual cycle removed) of precipitation, and zonal and meridional wind components at 850 hPa from the climate forecast system reanalysis (1979–2010). We examine associations between the intraseasonal variability of the A-ITCZ and the activity of the Madden–Julian oscillation (MJO). We show that during austral summer intraseasonal variability of the A-ITCZ is associated with a Rossby wave train in the Northern Hemisphere. In austral winter this variability is associated with the propagation of a Rossby wave in the Southern Hemisphere consistent with the Pacific-South American pattern. Moreover, we show that intense A-ITCZ events on intraseasonal time-scales are more frequent during the phase of MJO characterized by convection over western Pacific and suppression over the Indian Ocean. These teleconnection patterns induce anomalies in the trade winds and upper level divergence over the equatorial Atlantic that modulate the intensity of the A-ITCZ.

Spatial distribution and the interdecadal change of leading modes of heat budget of the mixed-layer in the tropical Pacific and the association with ENSO

Using heat budget diagnosis of ocean mixed layer from the Global Ocean Data Assimilation System, the spatial distribution of the leading modes of the heat budget was examined. The analysis was for the tropical Pacific in 1979–2013 and was based on combined empirical orthogonal function (CEOF) analysis. The interdecadal changes of the leading modes and their associations with El Nino-Southern Oscillation (ENSO) were also analyzed. The first leading CEOF mode (CEOF1) corresponds to the ENSO mature phase. The contribution from the zonal advection was relatively small along the equator, except the region near the Pacific coast of Central America. The vertical entrainment and diffusion (surface heat flux) had pronounced maxima with positive (negative) values along the equatorial central and eastern Pacific. The meridional advection displayed a different spatial pattern with large positive values on both sides of the equator and smaller values along the equator. The total meridional advection anomaly was mainly determined by advection of anomalous temperature by climatological current responsible for broadening of the ENSO SSTA pattern meridionally. The zonal advection varied almost simultaneously with the tendency of ocean temperature anomaly in the mixed layer. The second leading CEOF mode (CEOF2) included contribution to SSTA tendency during the ENSO developing phase. The distribution pattern and amplitude of the zonal advection in the eastern Pacific in CEOF2 was similar to but with opposite sign to that in CEOF1. The amplitudes of the other dynamical and thermodynamical terms were smaller than that in CEOF1 and spatial distributions displayed an opposite variation between the Pacific coast of Central America and central and eastern tropical Pacific in CEOF2. A comparison of two periods (1979–1999 and 2000–2013) suggested that coupling in the tropical Pacific weakened at ENSO time scales and shifted to a relatively higher frequency regime (from 2 to 4 years averaged in 1979–1999 to 1.5–3 years) after 2000.

Detection and Modeling of Vegetation Phenology Spatiotemporal Characteristics in the Middle Part of the Huai River Region in China

Vegetation plays an important role in atmospheric, hydrologic and biochemical cycles and is an important indicator of the impact of climate and human factors on the environment. In this paper, a method, which combines the empirical orthogonal function (EOF) and temporal unmixing analysis (TUA) methods, is applied to monitor the phenological characteristcs and spatial distribution of vegetation phenology in the middle part of the Huai River region. Based on the variance and EOF curves, the EOF provides the number of phenology modes, information which is the basis for an accurate temporal unmixing model. The TUA describes the temporal vegetation phenological details and spatial distribution. Importantly, this approach does not require assumptions, prior information or pre-defined thresholds. The vegetation phenology curves derived from the MODIS EVI data using the combined EOF and TUA methods display much more detail than the curves from Landsat TM using spectral mixture analysis (SMA). Additionally, the vegetation phenology spatial distribution from MODIS EVI is consistent with the field survey data. The combination method of EOF and TUA can be used to monitor vegetation phenology spatiotemporal change in a large area from time series of MODIS EVI data.

On the Relationship between the North Pacific Climate Variability and the Central Pacific El Niño

Abstract This study examined connections between the North Pacific climate variability and occurrence of the central Pacific (CP) El Niño for the period from 1950 to 2012. A composite analysis indicated that the relationship between the North Pacific sea surface temperature (SST), along with its overlying atmospheric circulation, and the CP El Niño during the developing and mature phases was changed when the occurrence frequency of the CP El Niño significantly increased after 1990. Empirical orthogonal function (EOF) and singular value decomposition (SVD) analyses of variability in the tropical Pacific and its relationship to the North Pacific show that the North Pacific anomalous SST and the atmospheric variability are more closely associated with the occurrence of the CP El Niño after 1990 than before 1990. There were noticeable differences in terms of the atmospheric variability conditions over the North Pacific, such as the North Pacific Oscillation (NPO)-like atmospheric variability during the spring and its associated SST anomalies during the following winter before 1990 and after 1990. In addition, combined EOF analysis also indicated that the NPO-like atmospheric circulation becomes more effective at playing a role in initiating El Niño after 1990. Consequently, such a change might have been associated with the frequent occurrence of the CP El Niño after 1990.

Spatiotemporal relations between water budget components and soil water content in a forested tributary catchment

AbstractWe examined 3 years of measured daily values of all major water budget components (precipitation P, potential evapotranspiration PET, actual evapotranspiration ET, and runoff R) and volumetric soil water content θ of a small, forested catchment located in the west of Germany. The spatial distribution of θ was determined from a wireless sensor network of 109 points with 3 measurement depths each; ET was calculated from eddy‐covariance tower measurements. The water budget was dominantly energy limited, with ET amounting to approximately 90% of PET, and a runoff ratio R/P of 56%. P, ET, and R closed the long‐term water budget with a residual of 2% of precipitation. On the daily time scale, the residual of the water budget was larger than on the annual time scale, and explained to a moderate extent by θ (R2 = 0.40). Wavelet analysis revealed subweekly time scales, presumably dominated by unaccounted fast‐turnover storage terms such as interception, as a major source of uncertainty in water balance closure. At weekly resolution, soil water content explained more than half (R2 = 0.62) of the residual. By means of combined empirical orthogonal function and cluster analysis, two slightly different spatial patterns of θ could be identified that were associated with mean θ values below and above 0.35 cm3/cm3, respectively. The timing of these patterns as well as the varying coherence between PET, ET, and soil water content responded to changes in water availability, including a moderate response to the European drought in spring 2011.

Relationship between Amazon biomass burning aerosols and rainfall over the La Plata Basin

Abstract. High aerosol loads are discharged into the atmosphere by biomass burning in the Amazon and central Brazil during the dry season. These particles can interact with clouds as cloud condensation nuclei (CCN) changing cloud microphysics and radiative properties and, thereby, affecting the radiative budget of the region. Furthermore, the biomass burning aerosols can be transported by the low-level jet (LLJ) to the La Plata Basin, where many mesoscale convective systems (MCS) are observed during spring and summer. This work proposes to investigate whether the aerosols from biomass burning may affect the MCS in terms of rainfall over the La Plata Basin during spring. Aerosol effects are very difficult to isolate because convective clouds are very sensitive to small environment disturbances; for that reason, detailed analyses using different techniques are used. The binplot, 2-D histograms and combined empirical orthogonal function (EOF) methods are used to identify certain environmental conditions with the possible effects of aerosol loading. Reanalysis 2, TRMM-3B42 and AERONET data are used from 1999 up to 2012 during September–December. The results show that there are two patterns associated with rainfall–aerosol interaction in the La Plata Basin: one in which the dynamic conditions are more important than aerosols to generation of rain; and a second one where the aerosol particles have a more important role in rain formation, acting mainly to suppress rainfall over the La Plata Basin. However, these results need further investigation to strengthen conclusions, especially because there are limitations and uncertainties in the methodology and data set used.

MJO structure associated with the higher-order CEOF modes

The real-time multivariate Madden–Julian oscillation (RMM; MJO) index has been widely employed to monitor the amplitude, phase, and time evolution of MJO events, as the index is formulated from the leading two combined-empirical orthogonal function (CEOF) modes of daily anomalous OLR and 850- and 200-hPa zonal winds, and the modes describe the MJO dynamics well. These two CEOF modes, however, are known to dominate in power spectra at zonal wavenumber one and may underestimate the power and structure at wavenumbers 2–5 where many MJO events are also prominent. This study approximated a baseline for MJO by applying band-pass filters to daily anomalies on 30–100 day periods and at 1–5 eastward propagating waves, as slightly different bands led to the same conclusions. Following the procedures to develop the RMM index, the daily anomalous data were derived and subjected to the CEOF analysis with all modes archived for diagnosis. Different numbers of the leading modes were compared in explained variance, standard deviation (STD), and wavenumber power spectra to describe the overall MJO magnitude and structure, and on the Hovmoller diagrams to represent the evolution of three distinct MJO events. Results show that the two leading CEOF modes explain only a small portion of the power spectra at wavenumbers 2–5. This spectral leakage notably reduces the MJO amplitude, particularly of the OLR in the western Pacific. The CEOF modes 3–10 can withhold power sufficiently such that the anomalies reconstructed by the first 10 modes contribute most of the baseline variance; their structures agree well with the baseline by constituting nearly the same proportion in the region from the central Indian Ocean to the dateline and by providing more complete evolutions of the three MJO events on the Hovmoller diagrams. Meanwhile, these modes introduce a notable amount of power for the equatorial Rossby and Kelvin waves that are partially embedded in the evolution of MJO. The first 50 of the total 432 CEOF modes retain all variance of the baseline MJO, while those higher than 10 contain less information and more noise and can be discarded. Furthermore, this study indicated that the longitudinal STD of the reconstructed anomalies detects the MJO phases and magnitudes in the western Pacific with more physical meaning and in better agreement with the Hovmoller diagrams than the RMM-like amplitude. The results provide an integral figure of the MJO structure from the CEOF analysis and a more robust RMM framework for monitoring the MJO’s evolution in real time and for validating its numerical forecast and simulations.

A Modified Multivariate Madden–Julian Oscillation Index Using Velocity Potential

Abstract A new Madden–Julian oscillation (MJO) index is developed from a combined empirical orthogonal function (EOF) analysis of meridionally averaged 200-hPa velocity potential (VP200), 200-hPa zonal wind (U200), and 850-hPa zonal wind (U850). Like the Wheeler–Hendon Real-time Multivariate MJO (RMM) index, which was developed in the same way except using outgoing longwave radiation (OLR) data instead of VP200, daily data are projected onto the leading pair of EOFs to produce the two-component index. This new index is called the velocity potential MJO (VPM) indices and its properties are quantitatively compared to RMM. Compared to the RMM index, the VPM index detects larger-amplitude MJO-associated signals during boreal summer. This includes a slightly stronger and more coherent modulation of Atlantic tropical cyclones. This result is attributed to the fact that velocity potential preferentially emphasizes the planetary-scale aspects of the divergent circulation, thereby spreading the convectively driven component of the MJO’s signal across the entire globe. VP200 thus deemphasizes the convective signal of the MJO over the Indian Ocean warm pool, where the OLR variability associated with the MJO is concentrated, and enhances the signal over the relatively drier longitudes of the equatorial Pacific and Atlantic. This work provides a useful framework for systematic analysis of the strengths and weaknesses of different MJO indices.

Eastward-Propagating Intraseasonal Oscillation Represented by Chikira–Sugiyama Cumulus Parameterization. Part I: Comparison with Observation and Reanalysis

Abstract The eastward-propagating intraseasonal oscillation represented by the Chikira–Sugiyama cumulus parameterization in an atmospheric general circulation model is compared with observations and reanalyses. The scheme is characterized by state-dependent entrainment rates that vertically vary affected by the environment. The zonal wavenumber–frequency power spectrum shows a strong signal corresponding to the Madden–Julian oscillation. The eastward-propagating feature of the convective region and accompanying anomalous zonal wind structure is well extracted by the first and second modes of the combined empirical orthogonal function (CEOF) with a reasonable explained variance for the first mode, though the second mode is not sufficiently reproduced. The basic features of the composited anomalous fields including moisture, temperature, and vertical and zonal winds resemble those of the reanalysis in both the free troposphere and surface air. The anomalous free-tropospheric moisture exhibits its westward tilt and the peak moist static energy of the surface air is shifted eastward as in the reanalysis. The anomalous low-level zonal wind changes its direction to the east of the convective center. The model’s outstanding deficiencies include the weak convective activity over the Indian Ocean, weaker westward tilt, and seemingly underestimated shallow convection.

Temporal and spatial characteristics of fog occurrence over the Korean Peninsula

In this study, fogs are classified based on the spatial and temporal characteristics over South Korea using the visibility data and the empirical orthogonal function (EOF) and wavelet analyses. With fog defined in terms of visibility (<1 km), the EOF analysis is performed to extract spatial distribution characteristics via dimension reduction, whereas the space‐time wavelet expansion is applied to the EOF time series to specify the fog characteristics in the space of time versus scale (i.e., period in this study). The first EOF mode occupies 48.9% of total variance and shows the fog distribution covering almost entire areas of South Korea with one sign (+), except at the eastern coast and western part of the southern coast. The wavelet analysis reveals that this fog occurs based on meteorological conditions of various scales from daily to seasonal, thus classified as mixed fog. The second EOF mode, which occupies 19.5% of total variance, shows distinct separation of spatial distribution of fog, with a negative (−) sign in winter over northwestern coastal/inland, western coastal, and south central mountain areas of South Korea and a positive (+) sign in other seasons elsewhere. With cycles of 1–2 weeks and 1–2 months being dominant in the wavelet analysis, this fog is considered to be strongly affected by synoptic scale weather systems and monsoon. Fog over the positive area is mostly affected by monsoon and/or cyclonic frontal systems, thus classified as frontal fog, whereas that over the negative area is affected by the cold‐core anticyclones moving over warm sea surface in winter or by radiative cooling, thus classified as steam fog (coastal/sea) or radiation fog (inland), respectively. The mountain area may have upslope fog because of orographic lifting. The third EOF mode, occupying 6.7% of total variance, depicts distinct spatial separation of fog distribution around the coastal areas with a negative (−) sign and in the inland areas with a positive (+) sign. The former, with a dominant 1–2 week cycle, is classified as sea fog affected by migratory anticyclones and monsoon in late spring and summer, while the latter, with a dominant diurnal variation, represents radiation fog under clear sky in autumn. It turns out that the combined EOF and wavelet analyses are useful to assess the detailed spatial and temporal characteristics of various types of fog occurrence in South Korea.

Empirical‐statistical downscaling of reanalysis data to high‐resolution air temperature and specific humidity above a glacier surface (Cordillera Blanca, Peru)

Recently initiated observation networks in the Cordillera Blanca (Peru) provide temporally high‐resolution, yet short‐term, atmospheric data. The aim of this study is to extend the existing time series into the past. We present an empirical‐statistical downscaling (ESD) model that links 6‐hourly National Centers for Environmental Prediction (NCEP)/National Center for Atmospheric Research (NCAR) reanalysis data to air temperature and specific humidity, measured at the tropical glacier Artesonraju (northern Cordillera Blanca). The ESD modeling procedure includes combined empirical orthogonal function and multiple regression analyses and a double cross‐validation scheme for model evaluation. Apart from the selection of predictor fields, the modeling procedure is automated and does not include subjective choices. We assess the ESD model sensitivity to the predictor choice using both single‐field and mixed‐field predictors. Statistical transfer functions are derived individually for different months and times of day. The forecast skill largely depends on month and time of day, ranging from 0 to 0.8. The mixed‐field predictors perform better than the single‐field predictors. The ESD model shows added value, at all time scales, against simpler reference models (e.g., the direct use of reanalysis grid point values). The ESD model forecast 1960–2008 clearly reflects interannual variability related to the El Niño/Southern Oscillation but is sensitive to the chosen predictor type.