Search Results
You are looking at 1 - 10 of 42 items for
- Author or Editor: Stefan Hastenrath x
- Refine by Access: Content accessible to me x
Abstract
General circulation mechanisms instrumental in both annual cycle and interannual variability of rainfall are studied with reference to key regions of the tropical Americas and Africa, including the Central AmericanāCaribbean area, northern Northeast Brazil, Subsaharan Africa, the Angola coast and the zaĆÆre (Congo) and Amazon basins. For most of these regions, rainfall anomalies tend to be associated with departures in the large-scale atmospheric and oceanic fields that correspond to the pattern changes in the annual alternation of dry and rainy seasons. The interannual variability of climate and circulation thus appears largely as enhancement and reduction of the annual cycle.
Abstract
General circulation mechanisms instrumental in both annual cycle and interannual variability of rainfall are studied with reference to key regions of the tropical Americas and Africa, including the Central AmericanāCaribbean area, northern Northeast Brazil, Subsaharan Africa, the Angola coast and the zaĆÆre (Congo) and Amazon basins. For most of these regions, rainfall anomalies tend to be associated with departures in the large-scale atmospheric and oceanic fields that correspond to the pattern changes in the annual alternation of dry and rainy seasons. The interannual variability of climate and circulation thus appears largely as enhancement and reduction of the annual cycle.
Abstract
A substantial portion of the interannual variability of rainfall at Jakarta, Java, can be predicted from antecedent pressure anomalies at Darwin, northern Australia; the pressure persistence, the concurrent correlation of pressure and rainfall, and the predictability of rainfall from antecedent pressure are all largest during the āeastā monsoon (June-November). Because of the relatively simple large-scale circulation setting, warranting a single predictor (Darwin pressure), this region is chosen for a series of experiments aimed at exploring the seasonality and secular variations of predictability, optimal length of dependent record, and updating of the regression base period used for predictions on the independent data set.
The major features of pressure-rainfall relationships are common through much of the 1911ā83 record, namely sign and general magnitude of correlations and the closer relationships during the east, as compared to the west monsoon. Considerable differences are, however, apparent between decades. Them may stem from both sampling deficiencies (noise) and real long-term changes of the pressure-rainfall couplings due to secular alterations in the large-scale circulation setting. The competition between these two factors is relevant concerning the optimal length of the dependent record used for predictions into the independent data set, as well as the updating of the regression base period.
Abstract
A substantial portion of the interannual variability of rainfall at Jakarta, Java, can be predicted from antecedent pressure anomalies at Darwin, northern Australia; the pressure persistence, the concurrent correlation of pressure and rainfall, and the predictability of rainfall from antecedent pressure are all largest during the āeastā monsoon (June-November). Because of the relatively simple large-scale circulation setting, warranting a single predictor (Darwin pressure), this region is chosen for a series of experiments aimed at exploring the seasonality and secular variations of predictability, optimal length of dependent record, and updating of the regression base period used for predictions on the independent data set.
The major features of pressure-rainfall relationships are common through much of the 1911ā83 record, namely sign and general magnitude of correlations and the closer relationships during the east, as compared to the west monsoon. Considerable differences are, however, apparent between decades. Them may stem from both sampling deficiencies (noise) and real long-term changes of the pressure-rainfall couplings due to secular alterations in the large-scale circulation setting. The competition between these two factors is relevant concerning the optimal length of the dependent record used for predictions into the independent data set, as well as the updating of the regression base period.
Abstract
A complex of anomalies in the premonsoon large-scale circulation setting heralds the interannual variability of India summer monsoon rainfall. The most prominent precursors of precipitation anomalies are the latitude position of the upper-air ridge over India, apparently reflecting the persistence of the boreal winter wind regime and its consequence for the establishment of the summer upper-air circulation; the temperature in southern Asia and the adjacent North Indian Ocean waters, a factor instrumental in heat-low development and hence the establishment of meridional pressure gradients and lower-tropospheric airstreams from the Southern Hemisphere; and indices of the Southern Oscillation, capturing pressure departure patterns spanning the global tropics. Stepwise multiple regression is used to extract from this āanomaly complexā the variance most pertinent to the interannual variability of Southwest monsoon rainfall, observations of pertinent elements being available for the period 1939ā81. Regression models developed on a portion of this record are then used to predict the summer monsoon rainfall anomalies of the years 1966ā81.
The correlations between the various precursors and the rainfall anomalies vary in the course of 1939ā81, being, on the whole, strongest in the 1950s and 1960s. While the April latitude position of the 500 mb ridge along 75Ā°E proves to be the strongest predictor, performance is improved by inclusion of other elements representing preseason temperature and the Southern Oscillation. Correlation, root-mean-square error, bias, and absolute error are used as measures of forecast performance. A set of experiments with the dependent dataset, ending in 1965, indicates that a regression base period of about 20 yr is optimal for predictions into the independent portion of the record. Another set of experiments, in which the regression base periods are successively updated to the year immediately preceding the year to be forecast, shows no improvement of predictions over the fixed regression base periods. āCross-validationā is not found less demanding than prediction proper. It is demonstrated that about half of the interannual variance in monsoon rainfall can be predicted from antecedent anomalies in the large-scale circulation setting.
Abstract
A complex of anomalies in the premonsoon large-scale circulation setting heralds the interannual variability of India summer monsoon rainfall. The most prominent precursors of precipitation anomalies are the latitude position of the upper-air ridge over India, apparently reflecting the persistence of the boreal winter wind regime and its consequence for the establishment of the summer upper-air circulation; the temperature in southern Asia and the adjacent North Indian Ocean waters, a factor instrumental in heat-low development and hence the establishment of meridional pressure gradients and lower-tropospheric airstreams from the Southern Hemisphere; and indices of the Southern Oscillation, capturing pressure departure patterns spanning the global tropics. Stepwise multiple regression is used to extract from this āanomaly complexā the variance most pertinent to the interannual variability of Southwest monsoon rainfall, observations of pertinent elements being available for the period 1939ā81. Regression models developed on a portion of this record are then used to predict the summer monsoon rainfall anomalies of the years 1966ā81.
The correlations between the various precursors and the rainfall anomalies vary in the course of 1939ā81, being, on the whole, strongest in the 1950s and 1960s. While the April latitude position of the 500 mb ridge along 75Ā°E proves to be the strongest predictor, performance is improved by inclusion of other elements representing preseason temperature and the Southern Oscillation. Correlation, root-mean-square error, bias, and absolute error are used as measures of forecast performance. A set of experiments with the dependent dataset, ending in 1965, indicates that a regression base period of about 20 yr is optimal for predictions into the independent portion of the record. Another set of experiments, in which the regression base periods are successively updated to the year immediately preceding the year to be forecast, shows no improvement of predictions over the fixed regression base periods. āCross-validationā is not found less demanding than prediction proper. It is demonstrated that about half of the interannual variance in monsoon rainfall can be predicted from antecedent anomalies in the large-scale circulation setting.
Abstract
Departure characteristics of the general circulation in the tropical Atlantic and eastern Pacific are studied in relation to extreme climatic events identified from collectives of long-term rainfall stations and other hydrometeorological parameters, with emphasis to the Central American-Caribbean region (CARIB). Ship observations during 1911ā72 compiled with a 1Ā° square latitude-longitude resolution and extending between 30Ā°Nā30Ā°S from the African coast to the eastern Pacific form a major observational basis.
The quality of the rainy season in CARIB has a large negative correlation with annual precipitation in the U.S. Central Great Plains, and with rainfall and sea surface temperature along the Ecuadorian/Peruvian littoral; in addition there are weak negative correlations with northeastern Brazil rainfall and with discharge and water level in northern tributaries of the Amazon, and a strong positive correlation with hydrometeorological events in Subsaharan Africa. Northeastern Brazil rainfall is strongly negatively correlated with sea surface temperature along the Eucador/Peru coast, but not with rainfall.
During extreme dry years in CARIB the North Atlantic high expands equatorward, meridional pressure gradients steepen, and the trades are stronger, albeit in a somewhat more southward location; at the same time, the ITCZ over the eastern Pacific stays farther south and the South Atlantic high contracts on its equatorward side. For extreme wet years, the reverse departure patterns from the 1911ā70 mean maps are characteristic. Sea surface temperature anomalies reflect a response to variations in the subtropical highs and major ocean currents: advection of cold waters in the eastern part of the oceans is favored by equatorward expansion of the subtropical high in the respective hemisphere; the wintertime Gulf Stream system has a distinct signature concomitant with departures in the equatorial and South Atlantic; and in the equatorial eastern Pacific departures reverse from the winter preceding toward the height of an extreme rainy season, warm waters in July/August being characteristic for drought in CARIB. In view of the strong spatial correlations, departure patterns constructed from stratification according to extreme events in CARIB are expected to have more general validity. Anomalous rainy season conditions are signaled in advance by large-scale departure patterns in January/February, thus offering the prospect of foreshadowing extreme rainy season behavior from the setup of low-latitude circulation during the preceding northern winter.
Abstract
Departure characteristics of the general circulation in the tropical Atlantic and eastern Pacific are studied in relation to extreme climatic events identified from collectives of long-term rainfall stations and other hydrometeorological parameters, with emphasis to the Central American-Caribbean region (CARIB). Ship observations during 1911ā72 compiled with a 1Ā° square latitude-longitude resolution and extending between 30Ā°Nā30Ā°S from the African coast to the eastern Pacific form a major observational basis.
The quality of the rainy season in CARIB has a large negative correlation with annual precipitation in the U.S. Central Great Plains, and with rainfall and sea surface temperature along the Ecuadorian/Peruvian littoral; in addition there are weak negative correlations with northeastern Brazil rainfall and with discharge and water level in northern tributaries of the Amazon, and a strong positive correlation with hydrometeorological events in Subsaharan Africa. Northeastern Brazil rainfall is strongly negatively correlated with sea surface temperature along the Eucador/Peru coast, but not with rainfall.
During extreme dry years in CARIB the North Atlantic high expands equatorward, meridional pressure gradients steepen, and the trades are stronger, albeit in a somewhat more southward location; at the same time, the ITCZ over the eastern Pacific stays farther south and the South Atlantic high contracts on its equatorward side. For extreme wet years, the reverse departure patterns from the 1911ā70 mean maps are characteristic. Sea surface temperature anomalies reflect a response to variations in the subtropical highs and major ocean currents: advection of cold waters in the eastern part of the oceans is favored by equatorward expansion of the subtropical high in the respective hemisphere; the wintertime Gulf Stream system has a distinct signature concomitant with departures in the equatorial and South Atlantic; and in the equatorial eastern Pacific departures reverse from the winter preceding toward the height of an extreme rainy season, warm waters in July/August being characteristic for drought in CARIB. In view of the strong spatial correlations, departure patterns constructed from stratification according to extreme events in CARIB are expected to have more general validity. Anomalous rainy season conditions are signaled in advance by large-scale departure patterns in January/February, thus offering the prospect of foreshadowing extreme rainy season behavior from the setup of low-latitude circulation during the preceding northern winter.
Abstract
Large-scale departure maps of sea level pressure (SLP) and sea surface temperature (SST) are presented for the tropical Atlantic and eastern Pacific Oceans, as obtained by stratification with respect to extreme climatic events in key regions of the tropical Americas. Drought in the Central American-Caribbean region is characterized by an equatorward expansion of the North Atlantic high, a band of anomalously cold water extending across the North Atlantic and a positive SST anomaly in the eastern Pacific. Drought in northeast Brazil is associated with high SLP over the South Atlantic and low SLP over the North Atlantic, cold water in the South Atlantic, a band of positive SST anomalies across the North Atlantic, and positive SST departures in the eastern Pacific. During the Ecuador/Peru El NinĢo, SLP in the eastern Pacific is low and SST high, and positive SLP departures dominate the tropical Atlantic.
Independently, preferred modes of departure configurations are identified from principal component analysis of SLP (1942ā71) and SST (1948ā71). The first four principal components of SLP explain 43, 26, 14 and 10%, and the first four SST components 42, 24, 10 and 6% of the variance. The first principal components of SLP and SST, and the sea temperature along the Ecuador/Peru coast are highly correlated. This ensemble of departure configurations closely replicates the ones characteristic of the Ecuador/Peru El NinĢo. The second principal components of SLP and SST are correlated, as is the third SLP with the fourth SST component. However, the departure patterns obtained by stratification with respect to regional climate anomalies provide no overall analogy to these pattern ensembles. The fourth principal SLP and the third SST components are highly correlated. Both possess a high correlation of one sign with rainfall in the Central American-Caribbean region, and a high correlation of opposite sign with precipitation in northeast Brazil. This pattern ensemble offers an excellent replication for the two sets of departure patterns obtained by stratification with respect to drought in the Central American-Caribbean region and north-east Brazil. An additional principal component analysis was performed in which SLP, SST, and the three aforementioned regional hydrometeorological time series simultaneously served as input. Results corroborate the separate SLP and SST principal component analyses.
The stratification and principal component analyses are complementary approaches, in that they yield realistic and physically plausible patterns. It is hypothesized that mass exchanges on the scale of the near-global tropics dominate the pressure pattern and are related to regional circulation changes and climate anomalies.
Abstract
Large-scale departure maps of sea level pressure (SLP) and sea surface temperature (SST) are presented for the tropical Atlantic and eastern Pacific Oceans, as obtained by stratification with respect to extreme climatic events in key regions of the tropical Americas. Drought in the Central American-Caribbean region is characterized by an equatorward expansion of the North Atlantic high, a band of anomalously cold water extending across the North Atlantic and a positive SST anomaly in the eastern Pacific. Drought in northeast Brazil is associated with high SLP over the South Atlantic and low SLP over the North Atlantic, cold water in the South Atlantic, a band of positive SST anomalies across the North Atlantic, and positive SST departures in the eastern Pacific. During the Ecuador/Peru El NinĢo, SLP in the eastern Pacific is low and SST high, and positive SLP departures dominate the tropical Atlantic.
Independently, preferred modes of departure configurations are identified from principal component analysis of SLP (1942ā71) and SST (1948ā71). The first four principal components of SLP explain 43, 26, 14 and 10%, and the first four SST components 42, 24, 10 and 6% of the variance. The first principal components of SLP and SST, and the sea temperature along the Ecuador/Peru coast are highly correlated. This ensemble of departure configurations closely replicates the ones characteristic of the Ecuador/Peru El NinĢo. The second principal components of SLP and SST are correlated, as is the third SLP with the fourth SST component. However, the departure patterns obtained by stratification with respect to regional climate anomalies provide no overall analogy to these pattern ensembles. The fourth principal SLP and the third SST components are highly correlated. Both possess a high correlation of one sign with rainfall in the Central American-Caribbean region, and a high correlation of opposite sign with precipitation in northeast Brazil. This pattern ensemble offers an excellent replication for the two sets of departure patterns obtained by stratification with respect to drought in the Central American-Caribbean region and north-east Brazil. An additional principal component analysis was performed in which SLP, SST, and the three aforementioned regional hydrometeorological time series simultaneously served as input. Results corroborate the separate SLP and SST principal component analyses.
The stratification and principal component analyses are complementary approaches, in that they yield realistic and physically plausible patterns. It is hypothesized that mass exchanges on the scale of the near-global tropics dominate the pressure pattern and are related to regional circulation changes and climate anomalies.
Abstract
The NCEPāNCAR 1958ā97 upper-air dataset and surface observations have been analyzed for evidence of zonalāvertical circulations along the Indian Ocean equator and their role in climatic variability. The long-term mean upper-tropospheric circulation is dominated in boreal winter by divergent outflow from the southern Indian Ocean northwestward into southern Asia, and in summer from southern Asia southwestward into the Southern Hemisphere. In boreal autumn only, divergent easterlies blow from Indonesia along the equator into an upper-tropospheric convergence band over East Africa, and only then a closed zonalāvertical circulation cell materializes along the Indian Ocean equator, between the centers of ascending motion over Indonesia and of subsidence over equatorial East Africa, and featuring westerlies in the lower layers. The boreal autumn zonalāvertical circulation varies interannually. A regime of intense circulation features accelerated equatorial surface westerlies, enhanced subsidence, and deficient rainfall at the coast of East Africa. In the high phase of the Southern Oscillation [anomalously high (low) pressure at Tahiti (Darwin)] this regime is preferred. The regime of weak zonal circulation has the opposite departure characteristics.
Abstract
The NCEPāNCAR 1958ā97 upper-air dataset and surface observations have been analyzed for evidence of zonalāvertical circulations along the Indian Ocean equator and their role in climatic variability. The long-term mean upper-tropospheric circulation is dominated in boreal winter by divergent outflow from the southern Indian Ocean northwestward into southern Asia, and in summer from southern Asia southwestward into the Southern Hemisphere. In boreal autumn only, divergent easterlies blow from Indonesia along the equator into an upper-tropospheric convergence band over East Africa, and only then a closed zonalāvertical circulation cell materializes along the Indian Ocean equator, between the centers of ascending motion over Indonesia and of subsidence over equatorial East Africa, and featuring westerlies in the lower layers. The boreal autumn zonalāvertical circulation varies interannually. A regime of intense circulation features accelerated equatorial surface westerlies, enhanced subsidence, and deficient rainfall at the coast of East Africa. In the high phase of the Southern Oscillation [anomalously high (low) pressure at Tahiti (Darwin)] this regime is preferred. The regime of weak zonal circulation has the opposite departure characteristics.
Abstract
A review is given of the current status of long-range forecasting in the low latitudes. Promising leads have developed over the past five years in the seasonal forecasting for certain target regions of the Tropics. Various approaches are of interest: (a) empirical methods based on the combination of general circulation diagnostics and statistical techniques; (b) numerical modeling, itself requiring also a diagnostic understanding from empirical analyses; and (c) purely statistical techniques. Regional targets include Indian monsoon, Nile and Ethiopia, eastern Africa, southern Africa, Sahel, Nordeste, North Atlantic storms, northwest Pacific storms, Australia, and El NiƱo and Southern Oscillation. Forecasts in real time are now being issued for most of these targets, and mostly from methods combining general circulation diagnostics with statistics. El NiƱo and Southern Oscillation are the target of real-time forecasting from purely statistical techniques, as well as the subject of numerical modeling by four different groups. Numerical modeling is also being applied in real-time climate prediction to Sahel and Nordeste.
From this experience it seems most fruitful to pursue concurrently all three lines of approach. Remarkable spontaneous initiatives in diagnostic monitoring and prediction have been undertaken in various parts of the world, including documentation of methods and verification of performance. Publication of real-time forecasts in a recently established forecast forum stimulates the professional interaction. The cultivation of the essential databases in real time, including further improvement of quality control and timely availability, may be the most important communal task at this stage.
Abstract
A review is given of the current status of long-range forecasting in the low latitudes. Promising leads have developed over the past five years in the seasonal forecasting for certain target regions of the Tropics. Various approaches are of interest: (a) empirical methods based on the combination of general circulation diagnostics and statistical techniques; (b) numerical modeling, itself requiring also a diagnostic understanding from empirical analyses; and (c) purely statistical techniques. Regional targets include Indian monsoon, Nile and Ethiopia, eastern Africa, southern Africa, Sahel, Nordeste, North Atlantic storms, northwest Pacific storms, Australia, and El NiƱo and Southern Oscillation. Forecasts in real time are now being issued for most of these targets, and mostly from methods combining general circulation diagnostics with statistics. El NiƱo and Southern Oscillation are the target of real-time forecasting from purely statistical techniques, as well as the subject of numerical modeling by four different groups. Numerical modeling is also being applied in real-time climate prediction to Sahel and Nordeste.
From this experience it seems most fruitful to pursue concurrently all three lines of approach. Remarkable spontaneous initiatives in diagnostic monitoring and prediction have been undertaken in various parts of the world, including documentation of methods and verification of performance. Publication of real-time forecasts in a recently established forecast forum stimulates the professional interaction. The cultivation of the essential databases in real time, including further improvement of quality control and timely availability, may be the most important communal task at this stage.
Abstract
The annual cycles of rainfall and river discharge in northern South America are dominated by the seasonal latitude migration of the intertropical convergence zone. The catchments in the west (Madden. Magdalena and Orinoco) have their high stands between August and December, while those of the Guyanas (Cuyuni, Essequibo, Suriname, Maroni and Oyapock) peak between May and July. Anomalously abundant discharge is in almost all catchments (except Orinoco) associated with the high SO phase (defined by anomalously high/low pressure at Tahiti/Darwin), weakened Caribbean tradewinds. and accelerated cross-equatorial southerly flow over the eastern Pacific.
In a series of experiments a sharp distinction was kept between a ādependentā dataset (1940ā70, or the part available in the river series) used as training period and an āindependentā portion of the record (1971ā87) reserved for prediction. Stepwise multiple regression models for bimonthly āseasonsā used as input river discharge as regressand, and as regressors index series of Tahiti minus Darwin pressure difference, equatorial Pacific sea surface temperature PWT, zonal wind component over the Caribbean and meridional wind component over the eastern equatorial Pacific, all two seasons earlier. The resulting equations were then used to predict the discharge anomalies in the independent dataset 1971ā87. There is considerable predictive skill for various rivers/seasons, with the overall best predictability for the low discharge time of year. In particular, for Magdalena 55% and for Essequibo 74% of the interannual variance of January-February discharge during 1971ā87 is predictable by this method, in which for Essequibo PWT serves as sole predictor.
Abstract
The annual cycles of rainfall and river discharge in northern South America are dominated by the seasonal latitude migration of the intertropical convergence zone. The catchments in the west (Madden. Magdalena and Orinoco) have their high stands between August and December, while those of the Guyanas (Cuyuni, Essequibo, Suriname, Maroni and Oyapock) peak between May and July. Anomalously abundant discharge is in almost all catchments (except Orinoco) associated with the high SO phase (defined by anomalously high/low pressure at Tahiti/Darwin), weakened Caribbean tradewinds. and accelerated cross-equatorial southerly flow over the eastern Pacific.
In a series of experiments a sharp distinction was kept between a ādependentā dataset (1940ā70, or the part available in the river series) used as training period and an āindependentā portion of the record (1971ā87) reserved for prediction. Stepwise multiple regression models for bimonthly āseasonsā used as input river discharge as regressand, and as regressors index series of Tahiti minus Darwin pressure difference, equatorial Pacific sea surface temperature PWT, zonal wind component over the Caribbean and meridional wind component over the eastern equatorial Pacific, all two seasons earlier. The resulting equations were then used to predict the discharge anomalies in the independent dataset 1971ā87. There is considerable predictive skill for various rivers/seasons, with the overall best predictability for the low discharge time of year. In particular, for Magdalena 55% and for Essequibo 74% of the interannual variance of January-February discharge during 1971ā87 is predictable by this method, in which for Essequibo PWT serves as sole predictor.
The term ādipoleā implies a seesaw, inverse variations of an element at the extremities of an oscillation system. The Southern Oscillation and North Atlantic Oscillation are examples of such mainly standing oscillations in pressure. By contrast, for the sea surface temperature (SST) fields in the tropical Atlantic and equatorial Indian Oceans, the SST gradient was found to be closely associated with climatic anomalies, but there is no seesaw. Use the term dipole is misleading.
The term ādipoleā implies a seesaw, inverse variations of an element at the extremities of an oscillation system. The Southern Oscillation and North Atlantic Oscillation are examples of such mainly standing oscillations in pressure. By contrast, for the sea surface temperature (SST) fields in the tropical Atlantic and equatorial Indian Oceans, the SST gradient was found to be closely associated with climatic anomalies, but there is no seesaw. Use the term dipole is misleading.
Climate-prediction research in the 1980s has shown particular promise for methods based on (a) general circulation and statistics, and (b) numerical modeling. Empirically based methods of predicting seasonal rainfall anomalies have been presented for India, Java, Kenya, Sahel, and Northeast Brazil. For some of these regions, about half of the interannual rainfall variability can be predicted from antecedent departures in the large-scale circulation. River discharge in northern South America, as well as atlantic tropical storm activity have proven highly predictable on empirical grounds. Numerical modeling has been used to advantage for the prediction of El NiƱo. Numerical modeling efforts are underway, directed to the forecasting of Sahel rainfall anomalies. Remarkable progress has been made towards the empirical prediction of food-grain production. A sound diagnostic understanding is crucial for the development of both empirical and numerical prediction methods. Among the most important tasks pending are the maintenance and timely processing of reliable, continuously functioning conventional raingauge networks; documentation of methods and verification of forecasts; and enhancement of contacts with the prospective users of climate prediction.
Climate-prediction research in the 1980s has shown particular promise for methods based on (a) general circulation and statistics, and (b) numerical modeling. Empirically based methods of predicting seasonal rainfall anomalies have been presented for India, Java, Kenya, Sahel, and Northeast Brazil. For some of these regions, about half of the interannual rainfall variability can be predicted from antecedent departures in the large-scale circulation. River discharge in northern South America, as well as atlantic tropical storm activity have proven highly predictable on empirical grounds. Numerical modeling has been used to advantage for the prediction of El NiƱo. Numerical modeling efforts are underway, directed to the forecasting of Sahel rainfall anomalies. Remarkable progress has been made towards the empirical prediction of food-grain production. A sound diagnostic understanding is crucial for the development of both empirical and numerical prediction methods. Among the most important tasks pending are the maintenance and timely processing of reliable, continuously functioning conventional raingauge networks; documentation of methods and verification of forecasts; and enhancement of contacts with the prospective users of climate prediction.