• Adeyewa, Z. D., , and K. Nakamura, 2003: Validation of TRMM radar rainfall data over major climatic regions in Africa. J. Appl. Meteor., 42, 331347.

    • Search Google Scholar
    • Export Citation
  • Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 11471167.

    • Search Google Scholar
    • Export Citation
  • Bolvin, D. T., , R. F. Adler, , G. J. Huffman, , E. J. Nelkin, , and J. P. Poutiainen, 2009: Comparison of GPCP monthly and daily precipitation estimates with high-latitude gauge observations. J. Appl. Meteor. Climatol., 48, 18431857.

    • Search Google Scholar
    • Export Citation
  • Boudreau, T., 2010: An Atlas of Ethiopian Livelihoods. U.S. Agency for International Development (USAID) and the government of Ethiopia, 138 pp.

    • Search Google Scholar
    • Export Citation
  • Bowden, J. H., , and F. H. M. Semazzi, 2007: Empirical analysis of intraseasonal climate variability over the greater Horn of Africa. J. Climate, 20, 57135731.

    • Search Google Scholar
    • Export Citation
  • Comrie, A. C., , and E. C. Glenn, 1998: Principal components-based regionalization of precipitation regimes across the southwest United States and northern Mexico, with an application to monsoon precipitation variability. Climate Res., 10, 201215.

    • Search Google Scholar
    • Export Citation
  • Conway, D., , E. Allison, , R. Felstead, , and M. Goulden, 2005: Rainfall variability in East Africa: Implications for natural resources, management and livelihoods. Philos. Trans. Roy. Soc. London, 363A, 4954.

    • Search Google Scholar
    • Export Citation
  • Conway, D., , A. Persechino, , S. Ardoin-Bardin, , H. Hamandawana, , C. Dieulin, , and G. Mahé, 2009: Rainfall and water resources variability in sub-Saharan Africa during the twentieth century. J. Hydrometeor., 10, 4159.

    • Search Google Scholar
    • Export Citation
  • Creswell, R., , and F. W. Martin, 1998: Dryland farming: Crops and techniques for arid regions. Educational Concerns for Hunger Organization (ECHO), N. Fort Meyers, FL, 23 pp.

    • Search Google Scholar
    • Export Citation
  • Dinku, T., , P. Ceccato, , E. Grover-Kopec, , M. Lemma, , S. J. Connor, , and C. F. Ropelewski, 2007: Validation of satellite rainfall products over East Africa’s complex topography. Int. J. Remote Sens., 28, 15031526.

    • Search Google Scholar
    • Export Citation
  • Dinku, T., , S. J. Connor, , and P. Ceccato, 2010: Comparison of CMORPH and TRMM-3B42 over mountainous regions of Africa and South America. Satellite Rainfall Applications for Surface Hydrology, M. Gebremichael and F. Hossaim, Eds., Springer, 193–204.

    • Search Google Scholar
    • Export Citation
  • Foeken, D., 1994: Climatic seasonality in Kenya with special reference to Coast province. East. South. Afr. Geogr. J., 5, 2843.

  • Frappart, F., and Coauthors, 2009: Rainfall regime across the Sahel band in the Gourma region, Mali. J. Hydrol., 375, 128142.

  • Gallo, K., , R. Hale, , D. Tarpley, , and Y. Yu, 2011: Evaluation of the relationship between air and land surface temperature under clear and cloudy sky conditions. J. Appl. Meteor. Climatol., 50, 767775.

    • Search Google Scholar
    • Export Citation
  • Garbutt, D. J., , R. D. Stem, , M. D. Dennett, , and J. Elston, 1981: A comparison of the rainfall climate of eleven places in West Africa using a two-part model for daily rainfall. Arch. Meteor. Geoph. Biokl. Ser. B, 29, 137155.

    • Search Google Scholar
    • Export Citation
  • Griffith, J. F., 1972: Climate of Africa. Vol. 10, World Survey of Climatology, Elsevier, 604 pp.

  • Grist, J. P., , and S. E. Nicholson, 2001: A study of the dynamic factors influencing the rainfall variability in the West African Sahel. J. Climate, 14, 13371359.

    • Search Google Scholar
    • Export Citation
  • Horn, L. H., , and R. A. Bryson, 1960: Harmonic analysis of the annual march of precipitation over the United States. Ann. Assoc. Amer. Geogr., 50, 157171.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 3855.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., , R. F. Adler, , D. T. Bolvin, , and G. Gu, 2009: Improving the global precipitation record: GPCP version 2.1. Geophys. Res. Lett., 36, L17808, doi:10.1029/2009GL040000.

    • Search Google Scholar
    • Export Citation
  • Hulme, M., , R. Doherty, , T. Ngara, , M. New, , and D. Lister, 2001: African climate change: 1900–2100. Climatic Change, 17, 145168.

  • Jia, Z., , and W. Luo, 2009: A modified climate diagram displaying net water requirements of wetlands in arid and semi-arid regions. Agric. Water Manage., 96, 13391343.

    • Search Google Scholar
    • Export Citation
  • Johnson, T. C., , and J. O. Malala, 2009: Lake Turkana and its link to the Nile. The Nile: Origin, Environments, Limnology and Human Use, H. J. Dumont, Ed., Springer, 297–304.

    • Search Google Scholar
    • Export Citation
  • Kamara, S., , T. Kuruppuarachchi, , E. R. Ranatunge, , Y. Hayashi, , M. Yokozawa, , M. Nishimori, , and T. Mikami, 2002: Multivariate statistical analysis of the seasonal rainfall regimes of the Guinea-Fouta Djallon Mountains of West Africa. J. Agric. Meteor., 58, 171183.

    • Search Google Scholar
    • Export Citation
  • Keen, C. S., , and P. D. Tyson, 1973: Seasonality of South African rainfall: A note on its regional delimitation using spectral analysis. Arch. Meteor. Geophys. Biokl., 21B, 207214.

    • Search Google Scholar
    • Export Citation
  • Kinuthia, J. H., 1992: Horizontal and vertical structure of the Lake Turkana jet. J. Appl. Meteor., 31, 12481274.

  • LeHouérou, H. N., 1996: Climate change, drought and desertification. J. Arid Environ., 34, 133185.

  • LeHouérou, H. N., 2010: Bioclimatology and Biogeography of Africa. Springer, 242 pp.

  • Mortimore, M. J., , and W. M. Adams, 2001: Farmer adaptation, change and “crisis” in the Sahel. Global Environ. Change, 11, 4957.

  • Mugalavai, E. M., , E. C. Kipkorir, , D. Raes, , and M. S. Rao, 2008: Analysis of rainfall onset, cessation and length of growing season for western Kenya. Agric. For. Meteor., 148, 11231135.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., 1996: A review of climate dynamics and climate variability in eastern Africa. Limnology, Climatology and Paleoclimatology of the East African Lakes, T. C. Johnson and E. O. Odada, Eds., CRC Press, 25–56.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., 1998: Historical fluctuations of Lake Victoria and other lakes in the Northern Rift Valley of East Africa. Environmental Change and Response in East African Lakes, J. T. Lehman, Ed., Kluwer, 7–35.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., 2000: The nature of rainfall variability over Africa on time scales of decades to millennia. Global Planet. Change, 26, 137158.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., , J. Y. Kim, , and J. Hoopingarner, 1988: Atlas of African Rainfall and Its Interannual Variability. The Florida State University, 237 pp.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., and Coauthors, 2003a: Validation of TRMM and other rainfall estimates with a high-density gauge dataset for West Africa. Part I: Validation of GPCC rainfall product and pre-TRMM satellite and blended products. J. Appl. Meteor., 42, 13371354.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., and Coauthors, 2003b: Validation of TRMM and other rainfall estimates with a high-density gauge dataset for West Africa. Part II: Validation of TRMM rainfall products. J. Appl. Meteor., 42, 13551368.

    • Search Google Scholar
    • Export Citation
  • Phillips, J., , and B. McIntyre, 2000: ENSO and interannual rainfall variability in Uganda: Implications for agricultural management. Int. J. Climatol., 20, 171182.

    • Search Google Scholar
    • Export Citation
  • Roca, R., , P. Chambon, , I. Jobard, , P.-E. Kirstetter, , M. Gosset, , and J. C. Bergès, 2010: Comparing satellite and surface rainfall products over West Africa at meteorologically relevant scales during the AMMA campaign using error estimates. J. Appl. Meteor. Climatol., 49, 715731.

    • Search Google Scholar
    • Export Citation
  • Todorov, A. V., , R. Achutuni, , and L. T. Steyaert, 1983: Agroclimatic conditions and assessment methods for drought/food shortages in the Horn of Africa. NOAA/NESDIS/Assessment and Information Services Center and the Atmospheric Science Department, University of Missouri—Columbia Rep., 89 pp.

    • Search Google Scholar
    • Export Citation
  • Vancutsem, C., , P. Ceccato, , T. Dinku, , and S. J. Connor, 2010: Evaluation of MODIS land surface temperature data to estimate air temperature in different ecosystems over Africa. Remote Sens. Environ., 114, 449465.

    • Search Google Scholar
    • Export Citation
  • Walter, H., , and H. Lieth, 1960: Klimadiagramm-Weltatlas (Climate Diagram World Atlas). Fischer-Verlag, 120 pp.

  • Wan, Z., 2008: New refinements and validation of the MODIS land-surface temperature/emissivity products. Remote Sens. Environ., 112, 5974.

    • Search Google Scholar
    • Export Citation
  • Wan, Z., , Y. Zhang, , Q. Zhang, , and Z.-l. Lia, 2002: Validation of the land-surface temperature products retrieved from Terra Moderate Resolution Imaging Spectroradiometer data. Remote Sens. Environ., 83, 163180.

    • Search Google Scholar
    • Export Citation
  • Wang, W., , S. Liang, , and T. Meyers, 2008: Validating MODIS land surface temperature products using long-term nighttime ground measurements. Remote Sens. Environ., 112, 623635.

    • Search Google Scholar
    • Export Citation
  • Zorita, E., , and F. F. Tilya, 2002: Rainfall variability in northern Tanzania in the March–May season (long rains) and its links to large-scale climate forcing. Climate Res., 20, 3140.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Seasonality map based on the TMPA at 0.25° spatial resolution, indicating nonseasonal as well as single- (1WS), dual- (2WS), and multiple-wet-season regimes and their modalities. Points A–D are reference locations used in a later figure.

  • View in gallery

    Seasonality map based on the GPCP at 2.5° spatial resolution. The legend and color scheme are the same as in Fig. 1, although not all of the classes appear in this map.

  • View in gallery

    Breakout map of East Africa from Fig. 1 to illustrate the spatial detail of the TMPA-based classification. Local names are indicated to show the correspondence of the map with local manifestations of wet seasons.

  • View in gallery

    Distribution of unimodal, bimodal, and trimodal rainfall regimes (reproduced from Nicholson et al. 1988, p. 9).

  • View in gallery

    Climate diagrams representative of different rainfall regimes for locations plotted in Fig. 1 as “A” (Mali: single wet season, unimodal), “B” (Tanzania: single wet season, bimodal), “C” (Somalia: dual wet season, unimodal-unimodal), and “D” (Ethiopia: multiple wet seasons) . The precipitation (P; bars) and temperature (2Ta: solid line; 4Ta: dotted line) plots determine the seasonality classes of each of these locations.

  • View in gallery

    Map of stability of rainfall seasonality regimes, portraying the percentage of individual years that fall into the same seasonality class as the mean of all 13 years on which the TMPA seasonality map in Fig. 1 is based.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 241 241 48
PDF Downloads 180 180 33

A Continental-Scale Classification of Rainfall Seasonality Regimes in Africa Based on Gridded Precipitation and Land Surface Temperature Products

View More View Less
  • 1 Hydrospheric and Biospheric Sciences Laboratory, NASA Goddard Space Flight Center, and Science Systems and Applications, Inc., Greenbelt, Maryland
  • 2 Laboratory for Atmospheres, NASA Goddard Space Flight Center, Greenbelt, Maryland
© Get Permissions
Full access

Abstract

A classification of rainfall seasonality regimes in Africa was derived from gridded rainfall and land surface temperature products. By adapting a method that goes back to Walter and Lieth’s approach of presenting climatic diagrams, relationships between estimated rainfall and temperature were used to determine the presence and pattern of humid, arid, and dry months. The temporal sequence of humid, arid, and dry months defined nonseasonal as well as single-, dual-, and multiple-wet-season regimes with one or more rainfall peaks per wet season. The use of gridded products resulted in a detailed, spatially continuous classification for the entire African continent at two different spatial resolutions, which compared well to local-scale studies based on station data. With its focus on rainfall patterns at fine spatial scales, this classification is complementary to coarser and more genetic classifications based on atmospheric driving forces. An analysis of the stability of the resulting seasonality regimes shows areas of relatively high year-to-year stability in the single-wet-season regimes and areas of lower year-to-year stability in the dual- and multiple-wet-season regimes as well as in transition zones.

Current affiliation: Office of Arid Land Studies, School of Natural Resources and the Environment, The University of Arizona, Tucson, Arizona.

Corresponding author address: Karen I. Mohr, Laboratory for Atmospheres, Code 613.1, NASA/GSFC, Greenbelt, MD 20771. E-mail: karen.mohr-1@nasa.gov

Abstract

A classification of rainfall seasonality regimes in Africa was derived from gridded rainfall and land surface temperature products. By adapting a method that goes back to Walter and Lieth’s approach of presenting climatic diagrams, relationships between estimated rainfall and temperature were used to determine the presence and pattern of humid, arid, and dry months. The temporal sequence of humid, arid, and dry months defined nonseasonal as well as single-, dual-, and multiple-wet-season regimes with one or more rainfall peaks per wet season. The use of gridded products resulted in a detailed, spatially continuous classification for the entire African continent at two different spatial resolutions, which compared well to local-scale studies based on station data. With its focus on rainfall patterns at fine spatial scales, this classification is complementary to coarser and more genetic classifications based on atmospheric driving forces. An analysis of the stability of the resulting seasonality regimes shows areas of relatively high year-to-year stability in the single-wet-season regimes and areas of lower year-to-year stability in the dual- and multiple-wet-season regimes as well as in transition zones.

Current affiliation: Office of Arid Land Studies, School of Natural Resources and the Environment, The University of Arizona, Tucson, Arizona.

Corresponding author address: Karen I. Mohr, Laboratory for Atmospheres, Code 613.1, NASA/GSFC, Greenbelt, MD 20771. E-mail: karen.mohr-1@nasa.gov

1. Introduction

The documented variety of rainfall regimes across the African continent complicates the characterization of continental-scale trends in both total annual rainfall and its seasonal distribution. A diverse array of seasonal controls is responsible for a mosaic of different rainfall regimes alternating at fine spatial scales—in particular, in East Africa (Nicholson 1998; Hulme et al. 2001; Conway et al. 2005, 2009). Evaluating the effects of climate change in these rainfall regimes requires a robust and spatially accurate representation of their seasonal climatology to track changes not only in total annual rainfall but also in the seasonal character of rainfall regimes, which is of crucial importance to rain-fed agriculture and pasture development in the semiarid and subhumid zones. There are a number of local- and regional-scale seasonality classifications (e.g., Keen and Tyson 1973; Garbutt et al. 1981; Foeken 1994; Nicholson 1996; Phillips and McIntyre 2000; Kamara et al. 2002; Zorita and Tilya 2002; Dinku et al. 2007; Mugalavai et al. 2008; Frappart et al. 2009), but they use different criteria for defining seasonality. Very few continental-scale classifications exist. Griffith (1972), Nicholson et al. (1988), and Nicholson (2000) produced continental-scale seasonality maps based on rain gauge data. Because in situ observations are sparse over wide stretches of the continent, a continental-scale analysis using only gauge data yields a map that has relatively coarse resolution.

The extensive data records of gridded, satellite-based rainfall estimates at a variety of spatial resolutions provide improved means for the continental-scale mapping of rainfall regimes, which thus far have been underutilized for this purpose. We therefore propose a classification of African rainfall regimes based on gridded rainfall and temperature products. These products have the advantage of spatial and temporal continuity, as compared with observations from synoptic weather stations, which tend to be sparse in space and of variable completeness in time. We adopt Walter and Lieth’s (1960) approach of relating monthly rainfall and temperature to define arid and humid months, which has been widely used with station data in bioclimatic analyses (e.g., LeHouérou 1996; Jia and Luo 2009). The use of gridded rainfall and temperature products lends a spatial dimension to this standard system for creating climate diagrams.

The choice of this method, as opposed to harmonic analysis or principal component analysis for a more genetic regionalization of rainfall seasonality (e.g., Horn and Bryson 1960; Comrie and Glenn 1998), stems from our motivation to characterize rainfall seasonality as experienced on the ground, in its role as a driver for land cover processes. This paper is hence descriptive and will not discuss the atmospheric processes behind different regimes, given that each region of Africa already has a relevant body of literature on those processes (e.g., Grist and Nicholson 2001; Bowden and Semazzi 2007). Our goal is to provide a basis for future continental-scale studies of the variability and trends in the characteristics of seasonal rainfall and their relationships with vegetation phenology in croplands and rangelands. In view of this goal, a fine-spatial-resolution product is desirable to match the resolution of available vegetation index data. Our product is also expected to be useful as a baseline against which shifts in rainfall regimes/seasonality over time can be compared.

2. Data and methods

a. Rainfall and temperature data products

Monthly accumulated rainfall estimates were obtained from the Tropical Rainfall Measuring Mission Multisatellite Precipitation Analysis (TMPA) 3B43 product at 0.25° spatial resolution for 1998–2010 (Huffman et al. 2007) and from the Global Precipitation Climatology Project (GPCP), version 2.1, product at 2.5° spatial resolution for the longer time period of 1979–2008 (Huffman et al. 2009). These precipitation products are derived from merged observations from satellite sensors and rain gauges. Rainfall-rate estimates from satellite sensors are based on proxy parameters related to radiative and emissive properties of cloud hydrometeors at visible, infrared, and microwave wavelengths (Adler et al. 2003; Bolvin et al. 2009).

The accuracy of rainfall products such as TMPA has been evaluated for West Africa (Nicholson et al. 2003a,b; Roca et al. 2010), East Africa (Dinku et al. 2007; Dinku et al. 2010), and Africa as a whole (Adeyewa and Nakamura 2003). Rainfall products were generally well correlated with rain gauge measurements at the monthly time scale, with the largest errors occurring in areas characterized by a low density of rain gauges—in particular, in mountainous regions such as the Ethiopian Highlands. Despite the uncertainties in limited areas, the TMPA and GPCP products appear to be well suited for use in a continental analysis of seasonality.

The Moderate Resolution Imaging Spectroradiometer (MODIS) monthly mean land surface temperature product (MOD11C3) at 0.05° spatial resolution for the period 2000–10 provided the daytime and nighttime monthly averages of land surface temperature used in this analysis. The MODIS land surface temperature (LST) estimate is derived from two thermal infrared (IR) band channels using a split-window algorithm that accounts for the intervening effects of atmospheric and surface emissivity (Wan et al. 2002; Wan 2008). The relationship between LST and 2-m air temperature is complex—in particular in daytime because solar radiation affects the IR signal. Validation studies of MODIS LST products suggest an underestimation of minimum air temperatures by nighttime LSTs of 3.3°–4.2°C for East Africa (Vancutsem et al. 2010). In contrast, an overestimation of maximum air temperatures by daytime LSTs ranged from less than 2°C in cloudy conditions to over 7°C in clear-sky conditions for selected stations in the United States (Wang et al. 2008; Gallo et al. 2011). We averaged mean monthly nighttime and mean monthly daytime LSTs, with the underestimation of minimum air temperatures and overestimation of maximum air temperatures reducing the overall estimation error. These monthly average LSTs served as useful proxies for monthly average air temperatures. For a small fraction of the grid cells, missing values in the MOD11C3 product occurred for several periods in time, which forced us to compute the mean values from a temporal subset for these grid cells. Because monthly average air temperature is relatively stable over a decade, we consider these data gaps to be negligible.

b. Defining seasonality regimes

In prior classifications of seasonality regimes, the terminology has been used interchangeably. For example, “bimodal” has been used to describe both a regime with two distinct wet seasons separated by a dry season (e.g., Foeken 1994; Conway et al. 2005) and a regime with one wet season and two peaks within this wet season (e.g., Nicholson et al. 1988). In the following, we use the terms “dual wet season” and bimodal to differentiate clearly between regimes with two wet seasons and regimes with one wet season containing two peaks.

From the precipitation products and air temperature proxies, we created a spatially explicit climatology of humid, arid, and dry months for the African continent, the temporal sequence of which determined the seasonality regime. The precipitation products were averaged into mean monthly rainfall over the respective periods of observation (13 yr for TMPA and 27 yr for GPCP). A similar procedure was performed for the air temperature proxies. The number of wet seasons was determined from the climatology by identifying arid and humid months for each grid cell, following a simple empirically derived relationship between monthly rainfall and temperature put forward by Walter and Lieth (1960). They define an arid month as one in which the precipitation P in millimeters is less than 2 times the temperature Ta in degrees Celsius (i.e., P < 2Ta). In a humid month, the precipitation exceeds 2 times the temperature (P > 2Ta). Because 2Ta has been found to be linearly related to annual potential evapotranspiration recorded from lysimeters across different climate zones, it can be used as a proxy for evapotranspiration, for which measured values are available only for a few stations (e.g., LeHouérou 1996, 2010).

The temporal sequence of arid and humid months was used to determine the seasonality regime of each grid cell. A seasonal regime is characterized by the presence of both humid and arid months. A wet season is a consecutive stretch of one or more humid months, and a dry season is a consecutive stretch of one or more arid months. A single-wet-season regime has only one wet season and one dry season in a year. In a dual-wet-season regime, two wet seasons alternate with two dry seasons. If more than two wet seasons separated by dry seasons are present, it is a multiple-wet-season regime.

The modality of a wet season is described by the number of rainfall peaks within this season. Several criteria for defining rainfall peaks, including absolute and relative rainfall thresholds, were tested and the results were compared with the local- and regional-scale studies of seasonality cited in section 1. A refinement of the temperature–precipitation relationship proved to be the most useful method for defining modality. In this scheme, dry months within each wet season are identified for which the precipitation is less than 4 times but greater than 2 times the temperature (2Ta < P < 4Ta). A dry month is thus less extreme than an arid month. For a monthly average temperature of 25°C, which is a realistic value for much of tropical Africa, 4Ta corresponds to the often-cited monthly rainfall threshold of 100 mm, below which water requirements for many crops are not satisfied and annual grasses with their short root systems suffer water stress (Todorov et al. 1983; Creswell and Martin 1998).

The presence and location of dry months within a wet season were used to determine rainfall peaks. Only one rainfall peak (i.e., no alternation of humid and dry months within the wet season) defines a unimodal wet season. A wet season with two rainfall peaks, separated by at least one dry month, is bimodal. More than two rainfall peaks separated by dry months make a wet season multimodal. The wet seasons of dual-wet-season regimes may have the same or different modalities. In multiple-wet-season regimes, the individual wet seasons are generally too short to have more than one peak.

3. Results and discussion

a. Spatial patterns of rainfall seasonality regimes

Figure 1 shows the map of seasonality classes using the TMPA, color coded by seasonality regime and modality. Although there are theoretically many possible wet-season and modality combinations, Fig. 1 depicts only the eight classes that could be identified from the spatiotemporal analysis of the TMPA. Seasonal rainfall regimes constitute more than 60% of the African continent. Nonseasonal arid regimes (about 30% of the continent) cover the Sahara and Namib Deserts, parts of the Kalahari and Karoo Deserts in southern Africa, and arid provinces in East Africa. Less than 5% of the continent is classified as nonseasonal humid, consisting of the central Congo Basin, eastern Madagascar, and small areas of the Guinea coast. Most of the seasonal rainfall regimes (>90%) are single-wet-season regimes. Dual-wet-season regimes are most common in East Africa, with smaller areas along the Guinea coast, in coastal Mozambique and South Africa, west of the Madagascar highlands, along the edges of the southern deserts, and around the Atlas Mountains. The predominant modality in both single- and dual-wet-season regimes is unimodal. Bimodal and multimodal wet seasons, as well as small areas of multiple-wet-season regimes, tend to be found in transitional zones between the single- and dual-wet-season regimes.

Fig. 1.
Fig. 1.

Seasonality map based on the TMPA at 0.25° spatial resolution, indicating nonseasonal as well as single- (1WS), dual- (2WS), and multiple-wet-season regimes and their modalities. Points A–D are reference locations used in a later figure.

Citation: Journal of Applied Meteorology and Climatology 50, 12; 10.1175/JAMC-D-11-024.1

For comparison, Fig. 2 shows the same seasonality classification but based on the coarser-resolution GPCP product. Both figures, displayed in the same color scheme, are broadly consistent with each other. This consistency is critical to establishing confidence in the classification derived from the shorter time series of the TMPA. Resampling the TMPA data to GPCP resolution resulted in a seasonality classification overlapping with the GPCP classification for the respective time period to 88%, with discrepancies at the peripheries of dual-wet-season and nonseasonal humid regimes (not shown). The coarser-resolution map (Fig. 2) lacks the fine mosaic of seasonality classes depicted in Fig. 1, in particular in East and South Africa. Two of the classes occurring in small pockets in the fine-resolution map (single-wet-season multimodal and dual-wet-season unimodal–bimodal) are not identified at all in the coarse-resolution map.

Fig. 2.
Fig. 2.

Seasonality map based on the GPCP at 2.5° spatial resolution. The legend and color scheme are the same as in Fig. 1, although not all of the classes appear in this map.

Citation: Journal of Applied Meteorology and Climatology 50, 12; 10.1175/JAMC-D-11-024.1

East Africa’s complex topography and location near the equator result in the continent’s richest mix of different seasonality classes in close spatial proximity. Figure 3 emphasizes the gain in detail from using the higher-resolution TMPA. The classification of East Africa compares well to previous local-scale studies of the dual-wet-season and single-wet-season bimodal regimes in Tanzania (Zorita and Tilya 2002), single-wet-season unimodal–multimodal regimes in Uganda (Phillips and McIntyre 2000), and the mix of single–multiple-wet-season regimes in southern Kenya (Foeken 1994; Mugalavai et al. 2008) and southern Ethiopia (Dinku et al. 2007). The arid Turkana region of northwestern Kenya stands out, consistent with the climatology of this region (Kinuthia 1992; Johnson and Malala 2009). The map also corresponds well to local characterizations of rainfall seasonality (Fig. 3), further corroborating the TMPA-based classification.

Fig. 3.
Fig. 3.

Breakout map of East Africa from Fig. 1 to illustrate the spatial detail of the TMPA-based classification. Local names are indicated to show the correspondence of the map with local manifestations of wet seasons.

Citation: Journal of Applied Meteorology and Climatology 50, 12; 10.1175/JAMC-D-11-024.1

It is important to note differences between this remote sensing–based classification and previous station-based continental-scale maps by Nicholson et al. (1988), reproduced in Fig. 4, and by Nicholson (2000). The most notable difference is that our remote sensing–based maps (Figs. 1 and 2) depict a larger proportion of the continent as being single-wet-season regime and a smaller proportion as being dual-wet-season regime than do previous classifications. Whereas the spatial resolution of the GPCP-based classification is roughly consistent with that of Nicholson’s map, the TMPA-based classification (Fig. 1) has finer spatial detail and more classes, with the largest differences visible in East Africa. The differences between this and previous classifications can be attributed to various factors. The classifications rely on different sets of input data, with station data being sparse in some regions, and cover different time periods. We use a method to derive seasonality that includes temperature data in addition to rainfall data. Our classification differentiates between the number of wet seasons and their modality. Nicholson et al. (1988) depict the number of rainfall peaks per year, which does not necessarily correspond to wet seasons, whereas Nicholson (2000) considers the number of wet seasons but not their modality. With no rainfall thresholds specified to define a rainfall peak, nonseasonal regimes do not occur in Nicholson et al. (1988), and, as a result, the Sahara Desert is classified in three different seasonal rainfall regimes.

Fig. 4.
Fig. 4.

Distribution of unimodal, bimodal, and trimodal rainfall regimes (reproduced from Nicholson et al. 1988, p. 9).

Citation: Journal of Applied Meteorology and Climatology 50, 12; 10.1175/JAMC-D-11-024.1

The climate diagrams in Fig. 5 show the typical annual course of precipitation and temperature for four grid cells (locations marked in Fig. 1) that represent different rainfall seasonality regimes. They are all located in savanna climate zones, with a mean annual rainfall between 600 and 1000 mm and agropastoral land use. The two temperature curves (2Ta and 4Ta; °C) and their intersections with the precipitation bars define humid, arid, and dry months. Point A represents a unimodal single-wet-season regime under the influence of the West African monsoon. In point B, the presence of dry months (January and February) creates two rainfall peaks in the Msimu wet season and makes it a bimodal single-wet-season regime. Point C illustrates a dual-wet-season regime with two short but intense wet seasons: the Gu and Deyr seasons (all labeled in Fig. 3). Three wet seasons, separated by arid months, are present in point D, of which the second two are weak. This location in southern Ethiopia represents a small transition zone between a dual-wet-season regime, in which the first season (Belg) is dominant with a second season (Keremt) occurring from October to November, and a bimodal single-wet-season regime, in which the second rainfall peak (Keremt) is dominant and occurs from June to September (Boudreau 2010). This complex regime was found to be variable in time, with class membership deviating from its average seasonality class approximately one-half of the time at this location.

Fig. 5.
Fig. 5.

Climate diagrams representative of different rainfall regimes for locations plotted in Fig. 1 as “A” (Mali: single wet season, unimodal), “B” (Tanzania: single wet season, bimodal), “C” (Somalia: dual wet season, unimodal-unimodal), and “D” (Ethiopia: multiple wet seasons) . The precipitation (P; bars) and temperature (2Ta: solid line; 4Ta: dotted line) plots determine the seasonality classes of each of these locations.

Citation: Journal of Applied Meteorology and Climatology 50, 12; 10.1175/JAMC-D-11-024.1

b. Temporal variability of rainfall seasonality regimes

Given the high year-to-year variability of rainfall and its temporal distribution, seasonality and modality in individual years may differ from the climatological seasonality regime as defined in the TMPA map (Fig. 1). Figure 6 illustrates the degree of variability by showing the percentage of individual years that fall into the same seasonality class as the classification of the 13-yr average, when applying the same criteria for defining seasonality to annual data. With mean monthly temperatures relatively stable from year to year, rainfall is the defining variable that causes interannual deviations from the expected seasonality regime. For grid cells in which the climatological seasonality regime definition is close to a threshold, small differences in rainfall in a particular month are sufficient to cause fluctuations between two or more regimes from year to year. The map hence includes noise from insignificant changes and uncertainties in the TMPA data but in its overall picture is nonetheless useful to show how stable/reliable the seasonality regimes are expected to be for particular areas. There is a gradient from nonseasonal arid regimes, which are the most stable (mean stability of 78%), over single-, dual-, and multiple-wet-season regimes, which are progressively less stable, to the nonseasonal humid regimes, the stability of which is fairly low (mean stability of 29%). The latter can be explained by rainfall dropping below the aridity threshold in at least 1 month of the year, making the regime seasonal by our definition, whereas longer-term averaging removes those irregular arid months. The regime stability map paints a picture that is very much in contrast to the coefficient of variability (i.e., the ratio of the standard deviation to the mean) of total annual rainfall (not shown), which tends to be highest in arid and lowest in humid areas. It also places some important agricultural regions into seasonal regimes of low stability. Given the biophysical relevance of the criteria used to define seasonality and modality with respect to vegetation water requirements, this translates into potential risks for the growing season. It is important to note, however, that such risks can also be present in areas that appear to be stable from a seasonality point of view, notably the dry margins of rain-fed agriculture such as the Sahel and northern Sudan zones (approximately 13.5°–15.5°N), where the single unimodal wet season is so short that it can hardly be interrupted by a dry or arid month and yet is significant enough not to disappear entirely. There, the risk of inadequate moisture supply is so permanent that agricultural systems have adapted to it by diversifying crops and livelihoods (Mortimore and Adams 2001).

Fig. 6.
Fig. 6.

Map of stability of rainfall seasonality regimes, portraying the percentage of individual years that fall into the same seasonality class as the mean of all 13 years on which the TMPA seasonality map in Fig. 1 is based.

Citation: Journal of Applied Meteorology and Climatology 50, 12; 10.1175/JAMC-D-11-024.1

A closer examination of the kind of seasonality regime changes reveals fluctuations in both directions. In some cases, a wet season or rainfall peak fails to appear in individual years. Cases in which an additional wet season or rainfall peak occurs are just as frequent, however. That is, longer wet seasons tend to get fragmented in individual years, either into two wet seasons by an arid month or into a bimodal wet season by a dry month. By region, the most frequent fluctuations are 1) in North Africa, from single wet season to dual wet season; 2) in the western Sahara, from nonseasonal arid to a single wet season; 3) along the Guinea coast, between a bimodal single-wet-season regime and a dual-wet-season regime; 4) in East Africa, from dual season to single season or multiple seasons; 5) in southern Africa, from single season to dual or multiple seasons; and 6) in central Africa, from nonseasonal humid to a single- or dual-wet-season regime. Whether more systematic transitions between climatological seasonality regimes are expected under projected climate change remains to be studied.

4. Summary and conclusions

This study developed a continental-scale seasonality classification for Africa from the MODIS land surface temperature product MOD11C3 and the TMPA (0.25°) and GPCP (2.5°) rainfall products. Following the method of Walter and Lieth (1960), the classification scheme used the combination of monthly average air temperature Ta and monthly mean daily rainfall P to define humid (P > 2Ta), arid (P < 2Ta), and dry (2Ta < P < 4Ta) months. Spatiotemporal analysis determined zones of single, dual, or multiple wet seasons from the pattern of humid and arid months and derived the modalities of the wet seasons from the presence of additional peaks separated by dry months.

In the TMPA-based classification, seasonal rainfall regimes were more than 60% of the continent, nonseasonal arid were approximately 30%, and nonseasonal humid were less than 5%. Of the seasonal regimes, 90% had a single wet season and were predominantly unimodal. Dual-wet-season regimes were most common in East Africa, with smaller areas in the Guinea coast, Madagascar, and northwestern and southern Africa. Bimodal and multimodal single and dual wet seasons and multiple-wet-season regimes could be found in transition zones between unimodal single and dual wet seasons. The interannual stability of the regime classifications was variable across space, with the highest stability in the in arid and single-wet-season regimes and the lowest stability in the multiple-wet-season and humid regimes as well as in transition zones.

This classification presents a valid picture of the seasonality experienced on the ground, as shown in local-scale studies that were based on rain gauge data. The GPCP-based classification was broadly consistent with the TMPA although lacking the detailed mix of seasonality classes associated with small transition zones. Both maps show that the Walter and Lieth (1960) approach of distinguishing arid and humid months provides a robust way for classifying seasonality across different climate zones that could be applied to other areas in which rainfall is the defining factor of seasonality. As compared with previous continental-scale classifications, the TMPA-based classification provided more spatial detail and a clearer distinction between number of wet seasons and their modality and hence could form a basis for evaluating the implications of climate change on rainfall seasonality regimes and on vegetation phenology within those regimes.

Acknowledgments

This work was funded by the NASA Precipitation Measuring Mission. The GPCP and TMPA products were developed by NASA/GSFC for the GEWEX Global Precipitation Climatology Project and the Tropical Rainfall Measuring Mission, respectively. The MODIS data products were obtained from the Warehouse Inventory Search Tool (WIST), maintained by the EOSDIS, NASA/GSFC. We consulted with George Huffman and David Bolvin (GSFC), Beth Mohr (Brandeis), and Andrew Comrie (The University of Arizona). We also acknowledge the three anonymous reviewers, whose comments helped to improve this manuscript.

REFERENCES

  • Adeyewa, Z. D., , and K. Nakamura, 2003: Validation of TRMM radar rainfall data over major climatic regions in Africa. J. Appl. Meteor., 42, 331347.

    • Search Google Scholar
    • Export Citation
  • Adler, R. F., and Coauthors, 2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 11471167.

    • Search Google Scholar
    • Export Citation
  • Bolvin, D. T., , R. F. Adler, , G. J. Huffman, , E. J. Nelkin, , and J. P. Poutiainen, 2009: Comparison of GPCP monthly and daily precipitation estimates with high-latitude gauge observations. J. Appl. Meteor. Climatol., 48, 18431857.

    • Search Google Scholar
    • Export Citation
  • Boudreau, T., 2010: An Atlas of Ethiopian Livelihoods. U.S. Agency for International Development (USAID) and the government of Ethiopia, 138 pp.

    • Search Google Scholar
    • Export Citation
  • Bowden, J. H., , and F. H. M. Semazzi, 2007: Empirical analysis of intraseasonal climate variability over the greater Horn of Africa. J. Climate, 20, 57135731.

    • Search Google Scholar
    • Export Citation
  • Comrie, A. C., , and E. C. Glenn, 1998: Principal components-based regionalization of precipitation regimes across the southwest United States and northern Mexico, with an application to monsoon precipitation variability. Climate Res., 10, 201215.

    • Search Google Scholar
    • Export Citation
  • Conway, D., , E. Allison, , R. Felstead, , and M. Goulden, 2005: Rainfall variability in East Africa: Implications for natural resources, management and livelihoods. Philos. Trans. Roy. Soc. London, 363A, 4954.

    • Search Google Scholar
    • Export Citation
  • Conway, D., , A. Persechino, , S. Ardoin-Bardin, , H. Hamandawana, , C. Dieulin, , and G. Mahé, 2009: Rainfall and water resources variability in sub-Saharan Africa during the twentieth century. J. Hydrometeor., 10, 4159.

    • Search Google Scholar
    • Export Citation
  • Creswell, R., , and F. W. Martin, 1998: Dryland farming: Crops and techniques for arid regions. Educational Concerns for Hunger Organization (ECHO), N. Fort Meyers, FL, 23 pp.

    • Search Google Scholar
    • Export Citation
  • Dinku, T., , P. Ceccato, , E. Grover-Kopec, , M. Lemma, , S. J. Connor, , and C. F. Ropelewski, 2007: Validation of satellite rainfall products over East Africa’s complex topography. Int. J. Remote Sens., 28, 15031526.

    • Search Google Scholar
    • Export Citation
  • Dinku, T., , S. J. Connor, , and P. Ceccato, 2010: Comparison of CMORPH and TRMM-3B42 over mountainous regions of Africa and South America. Satellite Rainfall Applications for Surface Hydrology, M. Gebremichael and F. Hossaim, Eds., Springer, 193–204.

    • Search Google Scholar
    • Export Citation
  • Foeken, D., 1994: Climatic seasonality in Kenya with special reference to Coast province. East. South. Afr. Geogr. J., 5, 2843.

  • Frappart, F., and Coauthors, 2009: Rainfall regime across the Sahel band in the Gourma region, Mali. J. Hydrol., 375, 128142.

  • Gallo, K., , R. Hale, , D. Tarpley, , and Y. Yu, 2011: Evaluation of the relationship between air and land surface temperature under clear and cloudy sky conditions. J. Appl. Meteor. Climatol., 50, 767775.

    • Search Google Scholar
    • Export Citation
  • Garbutt, D. J., , R. D. Stem, , M. D. Dennett, , and J. Elston, 1981: A comparison of the rainfall climate of eleven places in West Africa using a two-part model for daily rainfall. Arch. Meteor. Geoph. Biokl. Ser. B, 29, 137155.

    • Search Google Scholar
    • Export Citation
  • Griffith, J. F., 1972: Climate of Africa. Vol. 10, World Survey of Climatology, Elsevier, 604 pp.

  • Grist, J. P., , and S. E. Nicholson, 2001: A study of the dynamic factors influencing the rainfall variability in the West African Sahel. J. Climate, 14, 13371359.

    • Search Google Scholar
    • Export Citation
  • Horn, L. H., , and R. A. Bryson, 1960: Harmonic analysis of the annual march of precipitation over the United States. Ann. Assoc. Amer. Geogr., 50, 157171.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., and Coauthors, 2007: The TRMM Multisatellite Precipitation Analysis (TMPA): Quasi-global, multiyear, combined-sensor precipitation estimates at fine scales. J. Hydrometeor., 8, 3855.

    • Search Google Scholar
    • Export Citation
  • Huffman, G. J., , R. F. Adler, , D. T. Bolvin, , and G. Gu, 2009: Improving the global precipitation record: GPCP version 2.1. Geophys. Res. Lett., 36, L17808, doi:10.1029/2009GL040000.

    • Search Google Scholar
    • Export Citation
  • Hulme, M., , R. Doherty, , T. Ngara, , M. New, , and D. Lister, 2001: African climate change: 1900–2100. Climatic Change, 17, 145168.

  • Jia, Z., , and W. Luo, 2009: A modified climate diagram displaying net water requirements of wetlands in arid and semi-arid regions. Agric. Water Manage., 96, 13391343.

    • Search Google Scholar
    • Export Citation
  • Johnson, T. C., , and J. O. Malala, 2009: Lake Turkana and its link to the Nile. The Nile: Origin, Environments, Limnology and Human Use, H. J. Dumont, Ed., Springer, 297–304.

    • Search Google Scholar
    • Export Citation
  • Kamara, S., , T. Kuruppuarachchi, , E. R. Ranatunge, , Y. Hayashi, , M. Yokozawa, , M. Nishimori, , and T. Mikami, 2002: Multivariate statistical analysis of the seasonal rainfall regimes of the Guinea-Fouta Djallon Mountains of West Africa. J. Agric. Meteor., 58, 171183.

    • Search Google Scholar
    • Export Citation
  • Keen, C. S., , and P. D. Tyson, 1973: Seasonality of South African rainfall: A note on its regional delimitation using spectral analysis. Arch. Meteor. Geophys. Biokl., 21B, 207214.

    • Search Google Scholar
    • Export Citation
  • Kinuthia, J. H., 1992: Horizontal and vertical structure of the Lake Turkana jet. J. Appl. Meteor., 31, 12481274.

  • LeHouérou, H. N., 1996: Climate change, drought and desertification. J. Arid Environ., 34, 133185.

  • LeHouérou, H. N., 2010: Bioclimatology and Biogeography of Africa. Springer, 242 pp.

  • Mortimore, M. J., , and W. M. Adams, 2001: Farmer adaptation, change and “crisis” in the Sahel. Global Environ. Change, 11, 4957.

  • Mugalavai, E. M., , E. C. Kipkorir, , D. Raes, , and M. S. Rao, 2008: Analysis of rainfall onset, cessation and length of growing season for western Kenya. Agric. For. Meteor., 148, 11231135.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., 1996: A review of climate dynamics and climate variability in eastern Africa. Limnology, Climatology and Paleoclimatology of the East African Lakes, T. C. Johnson and E. O. Odada, Eds., CRC Press, 25–56.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., 1998: Historical fluctuations of Lake Victoria and other lakes in the Northern Rift Valley of East Africa. Environmental Change and Response in East African Lakes, J. T. Lehman, Ed., Kluwer, 7–35.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., 2000: The nature of rainfall variability over Africa on time scales of decades to millennia. Global Planet. Change, 26, 137158.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., , J. Y. Kim, , and J. Hoopingarner, 1988: Atlas of African Rainfall and Its Interannual Variability. The Florida State University, 237 pp.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., and Coauthors, 2003a: Validation of TRMM and other rainfall estimates with a high-density gauge dataset for West Africa. Part I: Validation of GPCC rainfall product and pre-TRMM satellite and blended products. J. Appl. Meteor., 42, 13371354.

    • Search Google Scholar
    • Export Citation
  • Nicholson, S. E., and Coauthors, 2003b: Validation of TRMM and other rainfall estimates with a high-density gauge dataset for West Africa. Part II: Validation of TRMM rainfall products. J. Appl. Meteor., 42, 13551368.

    • Search Google Scholar
    • Export Citation
  • Phillips, J., , and B. McIntyre, 2000: ENSO and interannual rainfall variability in Uganda: Implications for agricultural management. Int. J. Climatol., 20, 171182.

    • Search Google Scholar
    • Export Citation
  • Roca, R., , P. Chambon, , I. Jobard, , P.-E. Kirstetter, , M. Gosset, , and J. C. Bergès, 2010: Comparing satellite and surface rainfall products over West Africa at meteorologically relevant scales during the AMMA campaign using error estimates. J. Appl. Meteor. Climatol., 49, 715731.

    • Search Google Scholar
    • Export Citation
  • Todorov, A. V., , R. Achutuni, , and L. T. Steyaert, 1983: Agroclimatic conditions and assessment methods for drought/food shortages in the Horn of Africa. NOAA/NESDIS/Assessment and Information Services Center and the Atmospheric Science Department, University of Missouri—Columbia Rep., 89 pp.

    • Search Google Scholar
    • Export Citation
  • Vancutsem, C., , P. Ceccato, , T. Dinku, , and S. J. Connor, 2010: Evaluation of MODIS land surface temperature data to estimate air temperature in different ecosystems over Africa. Remote Sens. Environ., 114, 449465.

    • Search Google Scholar
    • Export Citation
  • Walter, H., , and H. Lieth, 1960: Klimadiagramm-Weltatlas (Climate Diagram World Atlas). Fischer-Verlag, 120 pp.

  • Wan, Z., 2008: New refinements and validation of the MODIS land-surface temperature/emissivity products. Remote Sens. Environ., 112, 5974.

    • Search Google Scholar
    • Export Citation
  • Wan, Z., , Y. Zhang, , Q. Zhang, , and Z.-l. Lia, 2002: Validation of the land-surface temperature products retrieved from Terra Moderate Resolution Imaging Spectroradiometer data. Remote Sens. Environ., 83, 163180.

    • Search Google Scholar
    • Export Citation
  • Wang, W., , S. Liang, , and T. Meyers, 2008: Validating MODIS land surface temperature products using long-term nighttime ground measurements. Remote Sens. Environ., 112, 623635.

    • Search Google Scholar
    • Export Citation
  • Zorita, E., , and F. F. Tilya, 2002: Rainfall variability in northern Tanzania in the March–May season (long rains) and its links to large-scale climate forcing. Climate Res., 20, 3140.

    • Search Google Scholar
    • Export Citation
Save