Heidke Skill Score Research Articles

Enhancement of Satellite Precipitation Estimation via Unsupervised Dimensionality Reduction

A methodology to enhance satellite precipitation estimation using unsupervised dimensionality reduction (UDR) techniques is developed. This enhanced technique is an extension to the precipitation estimation from remotely sensed imagery using an artificial neural network (PERSIANN) and cloud classification system (CCS) method (PERSIANN-CCS) enriched using wavelet features combined with dimensionality reduction. Cloud-top brightness temperature measurements from the Geostationary Operational Environmental Satellite (GOES)-12 are used for precipitation estimation at 4 km × 4 km spatial resolutions every 30 min. The study area in the continental U.S. covers parts of Louisiana, Arkansas, Kansas, Tennessee, Mississippi, and Alabama. Based on quantitative measures, root mean square error and Heidke skill score (HSS), the results show that the UDR techniques can improve the precipitation estimation accuracy. In addition, the independent component analysis is shown to have better performance than other UDR techniques; and in some cases, it achieves 10% improvement in the HSS.

Verification and usability of location specific medium range weather forecast for Kullu valley

Weather plays a crucial role in agriculture. Precipitation, temperature, humidity, wind speed and direction, drying conditions, dry and wet spells are the most important weather elements information about whom could play a significant role in farm planning and operations. Inclement weather events like drought and floods, cold and heat waves, hails, squalls, tropical storms severely affect the production. Occurrences of erratic weather are beyond human control. It is possible to adapt or mitigate their malevolent effect to some extend if the occurrence of the events is predicted well in advance and farmers are suitably advised to take ameliorative measures. Attempts were made to verify the weather forecasts received on every Tuesday and Friday from NCMRWF/IMD. The verification analysis was carried out on weekly, seasonal and annual basis using various verification techniques, viz., Ratio Score (RS), Critical Success Index (CSI), Heidke Skill Score (HSS), Hanssen and Kuipers Score (HK), Root Mean Square Error (RMSE), usability analysis and correlation approach during 2000-01 to 2009-10. The analysis depicted that ratio score on yearly basis was highest (74.6) during 2005-06 followed by 2004-05 (72.9) and 2003-04 (72.7). The value of H.K. score ranged between 24 and 42. The forecast found within quite usability range for most of the parameters but improvements are still possible. The correlation analysis showed that there was high correlation between observed and predicted values over the years. Hence, the forecast was found widely applicable among different user groups.

A New Integrated Weighted Model in SNOW-V10: Verification of Categorical Variables

This paper presents the verification results for nowcasts of seven categorical variables from an integrated weighted model (INTW) and the underlying numerical weather prediction (NWP) models. Nowcasting, or short range forecasting (0–6 h), over complex terrain with sufficient accuracy is highly desirable but a very challenging task. A weighting, evaluation, bias correction and integration system (WEBIS) for generating nowcasts by integrating NWP forecasts and high frequency observations was used during the Vancouver 2010 Olympic and Paralympic Winter Games as part of the Science of Nowcasting Olympic Weather for Vancouver 2010 (SNOW-V10) project. Forecast data from Canadian high-resolution deterministic NWP system with three nested grids (at 15-, 2.5- and 1-km horizontal grid-spacing) were selected as background gridded data for generating the integrated nowcasts. Seven forecast variables of temperature, relative humidity, wind speed, wind gust, visibility, ceiling and precipitation rate are treated as categorical variables for verifying the integrated weighted forecasts. By analyzing the verification of forecasts from INTW and the NWP models among 15 sites, the integrated weighted model was found to produce more accurate forecasts for the 7 selected forecast variables, regardless of location. This is based on the multi-categorical Heidke skill scores for the test period 12 February to 21 March 2010.

A linear merging methodology for high-resolution precipitation products using spatiotemporal regression

Currently, the only viable option for a global precipitation product is the merger of several precipitation products from different modalities. In this article, we develop a linear merging methodology based on spatiotemporal regression. Four highresolution precipitation products (HRPPs), obtained through methods including the Climate Prediction Center's Morphing (CMORPH), Geostationary Operational Environmental Satellite-Based Auto-Estimator (GOES-AE), GOES-Based Hydro-Estimator (GOES-HE) and Self-Calibrating Multivariate Precipitation Retrieval (SCAMPR) algorithms, are used in this study. The merged data are evaluated against the Arkansas Red Basin River Forecast Center's (ABRFC's) ground-based rainfall product. The evaluation is performed using the Heidke skill score (HSS) for four seasons, from summer 2007 to spring 2008, and for two different rainfall detection thresholds. It is shown that the merged data outperform all the other products in seven out of eight cases. A key innovation of this machine learning method is that only 6% of the validation data are used for the initial training. The sensitivity of the algorithm to location, distribution of training data, selection of input data sets and seasons is also analysed and presented.

TOWARD RELIABLE BENCHMARKING OF SOLAR FLARE FORECASTING METHODS

Solar flares occur in complex sunspot groups, but it remains unclear how the probability of producing a flare of a given magnitude relates to the characteristics of the sunspot group. Here, we use Geostationary Operational Environmental Satellite X-ray flares and McIntosh group classifications from solar cycles 21 and 22 to calculate average flare rates for each McIntosh class and use these to determine Poisson probabilities for different flare magnitudes. Forecast verification measures are studied to find optimum thresholds to convert Poisson flare probabilities into yes/no predictions of cycle 23 flares. A case is presented to adopt the true skill statistic (TSS) as a standard for forecast comparison over the commonly used Heidke skill score (HSS). In predicting flares over 24 hr, the maximum values of TSS achieved are 0.44 (C-class), 0.53 (M-class), 0.74 (X-class), 0.54 (>=M1.0), and 0.46 (>=C1.0). The maximum values of HSS are 0.38 (C-class), 0.27 (M-class), 0.14 (X-class), 0.28 (>=M1.0), and 0.41 (>=C1.0). These show that Poisson probabilities perform comparably to some more complex prediction systems, but the overall inaccuracy highlights the problem with using average values to represent flaring rate distributions.

A Comprehensive Assessment of CFS Seasonal Forecasts over the Tropics*

Abstract The 15-member ensemble hindcasts performed with the National Centers for Environmental Prediction Climate Forecast System (CFS) for the period 1981–2005, as well as real-time forecasts for the period 2006–09, are assessed for seasonal prediction skills over the tropics, from deterministic (anomaly correlation), categorical (Heidke skill score), and probabilistic (rank probability skill score) perspectives. Further, persistence, signal-to-noise ratio, and root-mean-square error analyses are also performed. The CFS demonstrates high skill in forecasting El Niño–Southern Oscillation (ENSO) related sea surface temperature (SST) anomalies during developing and mature phases, including that of different types of El Niño. During ENSO, the space–time evolution of anomalous SST, 850-hPa wind, and rainfall along the equatorial Pacific, as well as the mechanisms involved in the teleconnection to the tropical Indian Ocean, are also well represented. During ENSO phase transition and in the summer, the skill of forecasting Pacific SST anomalies is modest. An examination of CFS ability in forecasting seasonal rainfall anomalies over the U.S. Affiliated Pacific Islands (USAPI) indicates that forecasting the persistence of dryness from El Niño winter into the following spring/summer is skillful at leads > 3 months. During strong El Niño years the persistence is predicted by all members with a 6-month lead time. Also, the model is skillful in predicting regional rainfall responses during different types of El Niño. Since both deterministic and probabilistic skill scores converge, the suggestion is that the forecast is useful. The model’s skill in the real-time forecasts for the period 2006–09 is also discussed. The results suggest the feasibility that a dynamical-system-based seasonal prediction of precipitation over the USAPI can be considered.

Monthly Forecast of Indian Southwest Monsoon Rainfall Based on NCEP’s Coupled Forecast System

The monthly forecast of Indian monsoon rainfall during June to September is investigated by using the hindcast data sets of the National Centre for Environmental Prediction (NCEP)’s operational coupled model (known as the Climate Forecast System) for 25 years from 1981 to 2005 with 15 ensemble members each. The ensemble mean monthly rainfall over land region of India from CFS with one month lead forecast is underestimated during June to September. With respect to the inter-annual variability of monthly rainfall it is seen that the only significant correlation coefficients (CCs) are found to be for June forecast with May initial condition and September rainfall with August initial conditions. The CFS has got lowest skill for the month of August followed by that of July. Considering the lower skill of monthly forecast based on the ensemble mean, all 15 ensemble members are used separately for the preparation of probability forecast and different probability scores like Brier Score (BS), Brier Skill Score (BSS), Accuracy, Probability of Detection (POD), False Alarm Ratio (FAR), Threat Score (TS) and Heidke Skill Score (HSS) for all the three categories of forecasts (above normal, below normal and normal) have been calculated. In terms of the BS and BSS the skill of the monthly probability forecast in all the three categories are better than the climatology forecasts with positive BSS values except in case of normal forecast of June and July. The “TS”, “HSS” and other scores also provide useful probability forecast in case of CFS except the normal category of July forecast. Thus, it is seen that the monthly probability forecast based on NCEP CFS coupled model during the southwest monsoon season is very encouraging and is found to be very useful.

Lightning and Severe Weather: A Comparison between Total and Cloud-to-Ground Lightning Trends

Abstract Many studies over the past several decades have attempted to correlate trends in lightning (e.g., rates, polarity) to severe weather occurrence. These studies mainly used cloud-to-ground (CG) lightning information due to the ease of data availability, high detection efficiency, and broad coverage across the United States, with somewhat inconclusive results. Conversely, it has been demonstrated that trends in total lightning are more robustly correlated to severe weather occurrence, with rapid increases in total lightning observed 10s of minutes prior to the onset of severe weather. Unfortunately, total lightning observations are not as numerous, or available over the same areal coverage domain, as provided by CG networks. Relatively few studies have examined concurrent trends in both total and CG lightning within the same severe thunderstorm, or even large sets of thunderstorms using an objective lightning jump algorithm. Multiple studies have shown that the total flash rate rapidly increases prior to the onset of severe weather. What is untested within the same framework is the use of CG information to perform the same task. Herein, total and CG lightning trends for 711 thunderstorms occurring in four regions of the country were examined to demonstrate the increased utility that total lightning provides over CG lightning, specifically within the framework of developing a useful lightning-based severe weather warning decision support tool. Results indicate that while both lightning datasets demonstrate the presence of increased lightning activity prior to the onset of severe weather, the use of total lightning trends was more effective than CG trends [probability of detection (POD), 79% versus 66%; false alarm rate (FAR), 36% versus 53%; critical success index (CSI), 55% versus 38%; Heidke skill score (HSS), 0.71 versus 0.55]. Moreover, 40% of false alarms associated with total lightning, and 16% of false alarms with CG lightning trends, occurred when a lightning jump associated with a severe weather “warning” was already in effect. If these false alarms are removed, the FAR drops from 36% to 22% for total lightning and from 53% to 44% for CG lightning. Importantly, average lead times prior to severe weather occurrence were higher using total lightning as compared with CG lightning (20.65 versus 13.54 min). The ultimate goal of this study was to demonstrate the increased utility of total lightning information that the Geostationary Lightning Mapper (GLM) will provide to operational meteorology in anticipation of severe convective weather on a hemispheric scale once Geostationary Operational Environmental Satellite-R (GOES-R) is deployed in the next decade.

Another look at the contingency tables: scores based on Manhattan distances in the phase space

This alternative presentation of scores is based on the Manhattan distance in the phase space of forecasts. The key factor is represented by the ratio of the weights assigned to misses and false alarms. This ratio is 1 for the Heidke skill score and is equal to the ratio of the number of non-events to the number of events for the Pierce skill score. A score based on the deterministic limit leads to assign this ratio to either 2 or 0.5 as a function of the base rate. It can also be determined as a function of parameters of a cost-loss economical model. Applied to the Finley tornadoes and to two quantitative precipitation forecasts, the geographical representation of the phase space is quite invaluable as it points out the good points and the drawbacks of the forecast. The skill scores deduced from these distances help to summarize the quality of a forecast. Copyright © 2010 Royal Meteorological Society

A Statistical Forecast Model of Weather-Related Damage to a Major Electric Utility

AbstractA generalized linear model (GLM) has been developed to relate meteorological conditions to damages incurred by the outdoor electrical equipment of Public Service Electric and Gas, the largest public utility in New Jersey. Utilizing a perfect-prognosis approach, the model consists of equations derived from a backward-eliminated multiple-linear-regression analysis of observed electrical equipment damage as the predictand and corresponding surface observations from a variety of sources including local storm reports as the predictors. Weather modes, defined objectively by surface observations, provided stratification of the data and served to increase correlations between the predictand and predictors. The resulting regression equations produced coefficients of determination up to 0.855, with the lowest values for the heat and cold modes, and the highest values for the thunderstorm and mix modes. The appropriate GLM equations were applied to an independent dataset for model validation, and the GLM shows skill [i.e., Heidke skill score (HSS) values greater than 0] at predicting various thresholds of total accumulated equipment damage. The GLM shows higher HSS values relative to a climatological approach and a baseline regression model. Two case studies analyzed to critique model performance yielded insight into GLM shortcomings, with lightning information and wind duration being found to be important missing predictors under certain circumstances.

Influence of the Madden–Julian Oscillation on Forecasts of Extreme Precipitation in the Contiguous United States

Abstract Extreme precipitation events are among the most devastating weather phenomena since they are frequently accompanied by loss of life and property. This study uses reforecasts of the NCEP Climate Forecast System (CFS.v1) to evaluate the skill of nonprobabilistic and probabilistic forecasts of extreme precipitation in the contiguous United States (CONUS) during boreal winter for lead times up to two weeks. The CFS model realistically simulates the spatial patterns of extreme precipitation events over the CONUS, although the magnitudes of the extremes in the model are much larger than in the observations. Heidke skill scores (HSS) for forecasts of extreme precipitation at the 75th and 90th percentiles showed that the CFS model has good skill at week 1 and modest skill at week 2. Forecast skill is usually higher when the Madden–Julian oscillation (MJO) is active and has enhanced convection occurring over the Western Hemisphere, Africa, and/or the western Indian Ocean than in quiescent periods. HSS greater than 0.1 extends to lead times of up to two weeks in these situations. Approximately 10%–30% of the CONUS has HSS greater than 0.1 at lead times of 1–14 days when the MJO is active. Probabilistic forecasts for extreme precipitation events at the 75th percentile show improvements over climatology of 0%–40% at 1-day lead and 0%–5% at 7-day leads. The CFS has better skill in forecasting severe extremes (i.e., events exceeding the 90th percentile) at longer leads than moderate extremes (75th percentile). Improvements over climatology between 10% and 30% at leads of 3 days are observed over several areas across the CONUS—especially in California and in the Midwest.

QPF Model for lower Yamuna catchment, synoptic analogue method

The paper formulates a synoptic analogue model for issuing Quantitative Precipitation Forecast (QPF) for Lower Yamuna Catchment (LYC) based upon eleven years data (1998-2008) during southwest monsoon season. The results so derived were verified with realized Average Areal Precipitation (AAP) for the corresponding synoptic situation during 2009 southwest monsoon season. The performance of the model was observed Percentage Correct (PC) up to 86 % and for extreme events showed 100% correct with Heidke Skill Score (HSS) value 0.9. The experience during south west monsoon 2009 has shown that Synoptic analogue model can produce 24 hours advance QPF with accuracy and greater skill to facilitate the flood forecasters of Central Water Commission.

TESTING AUTOMATED SOLAR FLARE FORECASTING WITH 13 YEARS OF MICHELSON DOPPLER IMAGER MAGNETOGRAMS

Flare occurrence is statistically associated with changes in several characteristics of the line-of-sight magnetic field in solar active regions (ARs). We calculated magnetic measures throughout the disk passage of 1075 ARs spanning solar cycle 23 to find a statistical relationship between the solar magnetic field and flares. This expansive study of over 71,000 magnetograms and 6000 flares uses superposed epoch (SPE) analysis to investigate changes in several magnetic measures surrounding flares and ARs completely lacking associated flares. The results were used to seek any flare associated signatures with the capability to recover weak systematic signals with SPE analysis. SPE analysis is a method of combining large sets of data series in a manner that yields concise information. This is achieved by aligning the temporal location of a specified flare in each time series, then calculating the statistical moments of the “overlapping” data. The best-calculated parameter, the gradient-weighted inversion-line length (GWILL), combines the primary polarity inversion line (PIL) length and the gradient across it. Therefore, GWILL is sensitive to complex field structures via the length of the PIL and shearing via the gradient. GWILL shows an average 35% increase during the 40 hr prior to X-class flares, a 16% increase before M-class flares, and 17% increase prior to B–C-class flares. ARs not associated with flares tend to decrease in GWILL during their disk passage. Gilbert and Heidke skill scores are also calculated and show that even GWILL is not a reliable parameter for predicting solar flares in real time.

Convective energies in forecasting severe thunderstorms with one hidden layer neural net and variable learning rate back propagation algorithm

Thunderstorms are perennial features of India. However, the severe thunderstorms of pre — monsoon season (April–May) over Kolkata (22°32′N, 88°20′E) are of great concern for imparting devastating effect on life and property on the ground and aviation aloft. The study is thus, focused on developing one hidden layer neural network model with variable learning rate back propagation algorithm to forecast such thunderstorms. Convective available potential energy (CAPE) and convective inhibition energy (CIN) are selected as the input parameters of the model after the estimation of various skill scores like, Probability of Detection (POD), False Alarm Ratio (FAR), Heidke Skill Score (HSS) and Odds Ratio Skill Score (Yule’s Q) on different stability indices. During training the model, the squared error for forecasting severe thunderstorms is observed to be 0.0022 when the values of CIN within the range of 0 to 140 J kg−1 is taken as the input whereas the error is observed to be 0.0114 while the values of CAPE within the range of 2000 to 7000 J kg−1 is considered as the input. The values of CIN and CAPE at twelve to six hours prior to the occurrence of severe thunderstorms are considered in this study. During validation of the model, the percentage of prediction error with the values of CIN as input is observed to be 0.042% and that with CAPE as input is 0.162%. The values of CIN within the range of 0–140 J kg−1 are observed to be more persistent in forecasting severe thunderstorms over Kolkata than the values of CAPE within the range of 2000–7000 J kg−1.

Predicting potentially toxigenic Pseudo-nitzschia blooms in the Chesapeake Bay

Harmful algal blooms are now recognized as a significant threat to the Chesapeake Bay as they can severely compromise the economic viability of important recreational and commercial fisheries in the largest estuary of the United States. This study describes the development of empirical models for the potentially domoic acid-producing Pseudo - nitzschia species complex present in the Bay, developed from a 22-year time series of cell abundance and concurrent measurements of hydrographic and chemical properties. Using a logistic Generalized Linear Model (GLM) approach, model parameters and performance were compared over a range of Pseudo - nitzschia bloom thresholds relevant to toxin production by different species. Small-threshold blooms (≥10 cells mL − 1) are explained by time of year, location, and variability in surface values of phosphate, temperature, nitrate plus nitrite, and freshwater discharge. Medium- (100 cells mL − 1) to large- threshold (1000 cells mL − 1) blooms are further explained by salinity, silicic acid, dissolved organic carbon, and light attenuation (Secchi) depth. These predictors are similar to other models for Pseudo - nitzschia blooms on the west coast, suggesting commonalities across ecosystems. Hindcasts of bloom probabilities at a 19% bloom prediction point yield a Heidke Skill Score of ~53%, a Probability of Detection ∼ 75%, a False Alarm Ratio of ∼ 52%, and a Probability of False Detection ∼9%. The implication of possible future changes in Baywide nutrient stoichiometry on Pseudo - nitzschia blooms is discussed.

Evaluation of the National Hurricane Center’s Tropical Cyclone Wind Speed Probability Forecast Product

Abstract A tropical cyclone (TC) wind speed probability forecast product developed at the Cooperative Institute for Research in the Atmosphere (CIRA) and adopted by the National Hurricane Center (NHC) is evaluated for U.S. land-threatening and landfalling events over four hurricane seasons from 2004 to 2007. A key element of this work is the discernment of risk associated with the interval forecast probabilities for the three wind speed categories (i.e., 34, 50, and 64 kt, where 1 kt = 0.52 m s−1). A quantitative assessment of the interval probabilities (0–12, 12–24, 24–36, 36–48, 48–72, 72–96, and 96–120 h) is conducted by converting them into binary (yes–no) forecasts using decision thresholds that are selected using the true skill statistic (TSS) and the Heidke skill score (HSS). The NHC product performs well as both the HSS and TSS demonstrate skill out to the 48–72- and 72–120-h intervals, respectively. Overall, reliability diagrams and bias scores indicate that the NHC product has a tendency to overforecast event likelihood for cases where the forecast probabilities exceed 60%. Specifically, the NHC product tends to overforecast for the 34-kt category but underforecasts for the 64-kt category, especially at later forecast intervals. Results for the 50-kt category are mixed but also exhibit a tendency to underforecast during the latter intervals. Decision thresholds range from 1% to 55% depending on the selection method, wind speed category, and time interval. Given that the average forecast probabilities decrease with forecast hour, small forecast probabilities may be meaningful. The HSS is recommended over the TSS for decision threshold selection because the use of the TSS introduces significant bias and the HSS is less sensitive to filtering of correct negatives.

Comments on “Finite Samples and Uncertainty Estimates for Skill Measures for Seasonal Prediction”

Different skill measures present differing dependence on the intrinsic predictability level as well as differing dependence on sample size. Kumar (2009) computed the expected skill of idealized forecasts with varying levels of predictability using the anomaly correlation (AC), the Heidke skill score (HSS), and the ranked probability skill score (RPSS) as skill measures. An important consideration for seasonal climate forecasts, as demonstrated by Kumar (2009), is that these skill measures may vary significantly from their expected values when they are computed from relatively short forecast histories. In carrying out his skill measure characterizations, Kumar (2009) assumed that the forecast variance was equal to the climatological variance. Additionally, when computing average skill scores, forecast distributions were assumed to have identical means (signals) for a given predictability level. Here we examine some implications of those assumptions. This comment is organized as follows. In section 2 we present a perfect model framework for examining predictability. A natural requirement for predictability is that the forecast and climatological distributions be different. In the case of joint normal distributions, the existence of predictability (positive signal variance) implies that the forecast variance is less than the climatological variance. In this case, inflating the forecast variance to be equal to the climatological variance results in underconfident forecasts and lower probabilistic skill scores. In section 3 we compute the three skill scores for a given forecast signal (single initial condition) and for a set of forecasts with a specified signal to noise variance ratio (multiple initial conditions). The expected skill of a single forecast depends on both signal level and forecast variance. The AC depends on the ratio of squared signal to forecast variance, while the dependence of HSS and RPSS on signal level and forecast variance is more complex. A consequence of this functional dependence is that the AC of a set of forecasts is equal to that of a single forecast with the signal equal to the signal standard deviation. However, there is no similar relation for HSS and RPSS, and assuming a fixed signal as in Kumar (2009) generally overestimates HSS and always overestimates RPSS. We also present a useful approximation relating the RPSS and AC of a set of forecasts.

Using Fractal Downscaling of Satellite Precipitation Products for Hydrometeorological Applications

Abstract The objective of spatial downscaling strategies is to increase the information content of coarse datasets at smaller scales. In the case of quantitative precipitation estimation (QPE) for hydrological applications, the goal is to close the scale gap between the spatial resolution of coarse datasets (e.g., gridded satellite precipitation products at resolution L × L) and the high resolution (l × l; L ≫ l) necessary to capture the spatial features that determine spatial variability of water flows and water stores in the landscape. In essence, the downscaling process consists of weaving subgrid-scale heterogeneity over a desired range of wavelengths in the original field. The defining question is, which properties, statistical and otherwise, of the target field (the known observable at the desired spatial resolution) should be matched, with the caveat that downscaling methods be as a general as possible and therefore ideally without case-specific constraints and/or calibration requirements? Here, the attention is focused on two simple fractal downscaling methods using iterated functions systems (IFS) and fractal Brownian surfaces (FBS) that meet this requirement. The two methods were applied to disaggregate spatially 27 summertime convective storms in the central United States during 2007 at three consecutive times (1800, 2100, and 0000 UTC, thus 81 fields overall) from the Tropical Rainfall Measuring Mission (TRMM) version 6 (V6) 3B42 precipitation product (∼25-km grid spacing) to the same resolution as the NCEP stage IV products (∼4-km grid spacing). Results from bilinear interpolation are used as the control. A fundamental distinction between IFS and FBS is that the latter implies a distribution of downscaled fields and thus an ensemble solution, whereas the former provides a single solution. The downscaling effectiveness is assessed using fractal measures (the spectral exponent β, fractal dimension D, Hurst coefficient H, and roughness amplitude R) and traditional operational scores statistics scores [false alarm rate (FR), probability of detection (PD), threat score (TS), and Heidke skill score (HSS)], as well as bias and the root-mean-square error (RMSE). The results show that both IFS and FBS fractal interpolation perform well with regard to operational skill scores, and they meet the additional requirement of generating structurally consistent fields. Furthermore, confidence intervals can be directly generated from the FBS ensemble. The results were used to diagnose errors relevant for hydrometeorological applications, in particular a spatial displacement with characteristic length of at least 50 km (2500 km2) in the location of peak rainfall intensities for the cases studied.

A Physically Based Screen for Precipitation Over Complex Surfaces Using Passive Microwave Observations

Physically based passive microwave precipitation retrievals are difficult to develop over land because high nonuniform land emissivity values are difficult to distinguish from those of clouds. This paper uses an empirical approach to determine the covariance of emissivity at different microwave window channels and relies on this covariance to estimate the portion of the observed brightness temperatures that may be attributable to rainfall. One year (2006) of global cloud-free surface emissivity values were retrieved using data sets from multiple instruments on NASA's Aqua satellite. Correlations between the emissivities at different channels were developed for use in an empirical model within an optimal estimation retrieval. The optimal estimation retrieves surface temperature, total column water vapor, cloud water, and the emissivity at the 10.7-GHz horizontally polarized channel. From this retrieval and the covariance of emissivities, the 89.0-GHz brightness temperature at both polarizations can be estimated. Significant differences between the observed and retrieved high-resolution brightness temperatures are used to screen for precipitation, and results are compared to ground-based radar data for several study regions representing a variety of land surface types in the U.S. The Heidke Skill Score is used to determine the robustness of this methodology and, in all cases, demonstrates at least some increase in skill relative to random chance.

Techniques for evaluating the performance of landslide susceptibility models

Evaluating the performance of landslide susceptibility models is needed to ensure their reliable application to risk management and land-use planning. When results from multiple models are available, a comparison of their performance is necessary to select the model which performs better. In this paper, different techniques to evaluate model performance are discussed and tested using shallow landslide/debris-flow susceptibility models recently presented in the literature (Carrara, A., Crosta, G.B., Frattini, P., 2008. Comparing models of debris-flow susceptibility in the alpine environment. Geomorphology 94 (3–4), 353–378). Moreover, an evaluation technique based on the minimization of costs that may arise from the adoption of the model as a land management regulatory tool, is presented. The results of the application show that simple statistics such as Accuracy, Threat score, Gilbert's skill score, Pierce's skill score, Heidke skill score, and Yule's Q are problematic as they need to split the classified objects into two classes (e.g., stable/unstable) by defining an a-priori value of cutoff susceptibility, which is often not trivial. ROC curves and Success-Rate curves are cutoff-independent and can be used to efficiently visualize and compare the performance of models, but do not explicitly include classification costs. In addition, Success-Rate curves, under certain conditions, can be misleading when applied to grid-cell models. Cost curves include costs and a-priori probabilities, and are suitable for landslide susceptibility model performance evaluation from a practical point of view.