Changes in Wet Days and Dry Spells within the IGAD Region of Eastern Africa: An Observational Analysis


Mean wet days and dry spells

To facilitate the comparison of CHIRPS daily data with other 8 gridded Satellites derived from in-situ during the JJAS season in the IGAD region of eastern Africa used as results sample. The results show the highest number of wet days was recorded over the highlands of western Ethiopia and the western part of South Sudan (Fig. 2). These patterns of the highest wet days may explain the patterns of rainfall onset and cessation63 and the highest total rainfall amounts observed over South Sudan18, Kiremt rainy season over Ethiopia17 and Sudan64 during the JJAS season. More than 100 wets days were recorded over the highlands of western Ethiopia based on TAMSAT v3.1, NOAA-CPC RFEv2, PERSIANN-CDR, CMORPH RT V0.x BLD, TRMM 3B42 v7 and CPC v1.0. Out of 122 possible wet days during JJAS, the ASALs in Kenya, Somalia, southeastern Ethiopia and northern parts of Sudan recorded less than 15 wet days for all 9 gridded products. The CPC v1.0 has fewer wet days over Uganda, South Sudan and Ethiopia compared to other products. Even though JJAS is the dry season in most parts of Kenya, the highlands of western and Nyanza counties had more than 50 wet days. This indicates a high likelihood of success for rain-fed agricultural activities throughout the year considering the patterns of MAM and OND as the peak of the Long Rains and Short Rains over Kenya15.

Figure 2
figure 2

Comparison of JJAS means state of wet days of 9 gridded satellite datasets derived from the in-situ seasonal mean state of Number of wet days patterns(days/season) at 1 mm thresholds for the period 1981–2010 reference period. The pixel values in the legend are presented as the number of wet days.

Figure 3 shows the average number of wet days during MAM, JJAS, OND and DJF seasons based on CHIRPS datasets. Generally, the results show the areas with the highest number of wet days vary from season to season and within countries. On a seasonal timescale, the MAM season, the highest number of wet days (55–65 days) are recorded over southwestern South Sudan, southern and south-Eastern parts of Uganda, Lake Victoria basin and the highland of western and Nyanza region of Kenya. Findings by[41]found that the seasonal rainfall totals and the number of wet days at the sub-regional level in Equatorial Eastern Africa had the highest level of spatial coherence. The Majority of districts in Sudan, Eritrea, Djibouti, northern Somalia, Upper Nile state in South Sudan, southeastern, northern and northeastern Ethiopia, and northeastern Kenya recorded the lowest number of wet days (Fig. 3a). Similarly, the highest number of wet days (80–100) during JJAS were recorded over western South Sudan, the highlands of western Ethiopia, while northern Sudan, southeastern Ethiopia zones, the majority of counties in Kenya, most parts of districts in Somalia and coastal Djibouti recorded the lowest number of wet days (Fig. 3b). In OND season, the highest number of wet days were recorded over most parts of Uganda and Nyanza counties in Kenya. Most parts of Sudan, eastern and southeastern Ethiopia, Kenya, Somalia, Djibouti, and Eritrea recorded the lowest number of wet days ranging between 45 and 60 dry days (Fig. 3c). During the DJF season, with the exception of southern and central Uganda, Lake Victoria basin, the rest of IGAD regions recorded the lowest wet days ranging between 0 and 10 dry days (Fig. 3d).

Figure 3
figure 3

Seasonal mean state of Number of wet days patterns(days/season) at 1 mm thresholds for the period 1981–2010 reference period during (a) MAM, (b) JJAS, (c) OND and (d) DJF. The pixel values in the legend are presented as a number of wet days.

Figure 4 presents the comparison of mean dry spells(climatology) patterns of 9 gridded satellites derived from in-situ during JJAS seasons. Except CPC v1.0, other datasets agreed on 0–1 spells as the shortest period, which was observed over southwestern South Sudan, the highlands of central and western Ethiopia. The longest continued dry-spells recorded over ASALs in Kenya, Somalia, southeastern Kenya, Djibouti, Eritrea and squalled line in central Sudan.

Figure 4
figure 4

Comparison of JJAS mean state of wet days of 9 gridded Satellite derived from in-situ seasonal mean state of Number of dry spells patterns(days/season) at 1 mm thresholds for the period 1981–2010 reference period. The pixel values in the legend are presented as a number of dry spells.

Figure 5 presents the mean dry spells(climatology) patterns during four seasons. Conversely, the shortest and longest continued dry-spells recorded over areas with the highest and lowest wet days as described in Fig. 4 above. Most parts of Kenya, Somalia and Ethiopia recorded 3–5 consecutive dry spells (21–35 consecutive dry days) during MAM (Fig. 5a). The southwestern and northern parts of South Sudan, the highland of western Ethiopia, and rain belts in southern parts of Sudan recorded less than 1 spell as the shortest spell in the IGAD region during the JJAS season (Fig. 5b). Nevertheless, OND season shows 3–5 spells (21–35 consecutive dry days) recorded over most parts of Kenya, Somalia and southeastern Ethiopia. The majority of districts in southwestern Uganda recorded 1 spell as the longest (Fig. 5c). In a related study, an increase in the length of maximum dry spells was reported across the majority of locations in April and May over Uganda’s Lake Kyoga Basin28. Similarly, central and northern Uganda, most parts of Kenya, northeastern, central and southwestern Ethiopia have the longest consecutive spells during DJF season (Fig. 5d). Lake Victoria basin consistently observed lowest the spells across all four seasons.

Figure 5
figure 5

Seasonal mean state of Number of wet days patterns(days/season) at 1 mm thresholds for the period 1981–2010 reference period during (a) MAM, (b) JJAS, (c) OND and (d) DJF. The pixel values in the legend are presented as numbers of dry spells.

The variation in an average number of wet days and dry spells over Arua, AlQadarif, Upper Nile, Arsi, and Trans Nzoia is presented in Fig. 6. The variation assessed during the 1980s, 1990s, 2000s, and 2010s (1981–1990, 1991–2000, 2001–2010, 2011–2020), 30 years’ average (1981–2010), 41 years’ average (1981–2021) and 20 years’ average (1981–2000, 2001–2020). The results revealed that the mean values for four decades (1980s, 1990s, 2000s, 2010s) and 41 years average (1981–2021) for MAM, JJAS and OND Seasons recorded very close mean values across all five regions. The different timescale means values for the MAM season showed that Arua in Uganda recorded the highest mean of wet days (lowest mean of consecutive dry spells). AlQadarif in Sudan recorded the lowest mean wet days (highest mean dry spells) during the season (Fig. 6a,b). The highest wet days (lowest dry spells) during JJAS were recorded over Arua and Upper Nile in South Sudan (Fig. 6c,d). The Arua recorded the highest mean of wet days (lowest mean consecutive dry spells) during JJAS and continued to OND (Fig. 6e,f). in Addition, the Arsi zone in Ethiopia recorded the lowest mean values of wet days (highest mean value of consecutive dry spells) during OND. The fluctuation of the Madden Julian Oscillation during the MAM and OND seasons has been connected to consecutive wet days and dry spells periods over most parts of East Africa65.

Figure 6
figure 6

Histogram illustrate wet days (first column) and dry spells (second column) means in last 41 years (1981–2021), four decades (1981–1990, 1991–2000, 2001–2010 and 2011–2020), current 20 year (2001–2020) and previous 20 years (1981–2000) over El Gadaref in Sudan, El renk in South Sudan, Arsi in Ethiopia, Trans Nzoia in Kenya, Arua in Uganda.

Probability of wet days and dry spells exceeding defined thresholds

Figure 7 shows the probability of wet days exceeding 7, 14, 21, 28 consecutive wet days during MAM, JJAS, OND and DJF seasons. The results for the MAM season show that Uganda, South Sudan and Ethiopia, Kenya, Somalia, Djibouti and most parts of Eritrea observed more than 90% probability of wet days exceeding 7 consecutive wet days (Fig. 7a). However, the probability started shrinking continuously when 14, 21 and 28 consecutive wet days were examined. Southwestern South Sudan and Ethiopia, Uganda, and Nyanza countries in Kenya maintained more than 80% probability of exceeding 7, 14, 21 and 28 consecutive wet days during MAM (Fig. 7a–d) and JJAS seasons (Fig. 7e–h). Similarly, Lupi reported that in the Melkassa location in Central Rift Valley of Oromia State, Ethiopia, the likelihood of getting dry spells lasting 5, 7, and 10 days is less than 50%, and it lowers to under 20% at the start of the JJA peak season66. ASAL of Sudan, Kenya and Somalia recorded less than 1% probability of wet days exceeding 14, 21, and 28 consecutive days. Most parts of Uganda, South Sudan, Lake Victoria basin recorded more than 80% of wet days to exceed 14, 21, and 28 consecutive days across MAM, JJAS and OND season (Fig. 7i–l). Most parts of the IGAD region recorded less than 1% probability of wet days to exceed 14, 21, and 28 consecutive days during the DJF season (Fig. 7m–p). No changes in the pattern of the probability of exceeding 7–28 days over northern parts of Sudan across all four seasons (MAM, JJAS, OND and DJF) because these areas are dry climatologically.

Figure 7
figure 7

Probability of exceeding 7, 14, 21, 28 consecutive wet days at 1 mm rainfall thresholds during MAM (first raw), JJAS (second raw), OND (third raw) and DJF (fourth raw) season reference to 1981–2010 mean. For the period 1981–2010 reference period.

On the other hand, the probability of dry spells exceeds 1, 2, 3, 4 consecutive spells (7, 14, 21 and 28 days without rain) during MAM, JJAS, OND and DJF season presented in Fig. 8. The results show, the areas with lowest probability consecutive wet days (less than a 1%) discussed in Fig. 7 above recorded the highest (100%) probability of exceeding 1, 2, 3 and 4 consecutive dry spells (7, 14, 21, 28 days without rain) during MAM, JJAS, OND and DJF seasons (Fig. 8a–q). The patterns of dry spells in MAM over Eastern Uganda(70–90%) are different from the findings by Ojara et al., who reported that the probability of 8 days dry spell is high (38–69%) across all 9 weather stations over Lake Kyoga Basin in Uganda during in March, April, and August28. The difference is due to the usage of 7 and 8 days, and CHIRPS data appears to overestimate dry days compared to station data. Sudan, Eritrea, Ethiopia, Kenya, Djibouti and Somalia recorded a 100% probability of exceeding 1 consecutive dry spell (7 days without rain) during MAM, OND and DJF seasons (Fig. 8a,i,m). The patterns over Western South Sudan, southern parts of Sudan, and highlands of western Ethiopia are the only areas that recorded the lowest probability (less than 5%) of the probability of exceeding1 consecutive dry spell during JJAS season (Fig. 8e). Our finding is partially in agreement with study by67, who found the probability of three-dekad dry spells(30 days) was high at all stations in the Upper Awash River Basin, Ethiopia during the short rainy (Belg) season.

Figure 8
figure 8

Probability of exceeding 1, 2, 3, 4 consecutive spells during at 1 mm rainfall thresholds during MAM (first raw), JJAS (second raw), OND (third raw) and DJF (fourth raw) season reference to 1981–2010 mean. For the period 1981–2010 reference period.

Observed change in wet days and dry spells patterns

Figure 9 presents the observed rate of change (%) in the seasonal mean state of the number of wet days at 1 mm thresholds for the four decades 1980s, 1990s, 2000s, and 2010s. The results show a considerable decadal (10-year) fluctuation in wet days across the IGAD region. During the recent decade (2011–2020), nearly all parts of the region experienced an exceptional (unusual) increased number of wet days during the MAM, JJAS, and OND seasons. In the 1980s, wet days increased by more than 35% throughout Kenya, southern parts of Somalia, southeast Ethiopia, western Ethiopia, and southern portions of Sudan. In the southern regions of Sudan during the JJAS season in the 1980s and 1990s, there was a 15–30% increase in rainy days in the 2010s. Similarly, during the OND season in the 1990s, 2000s, and 2010, most parts of South Sudan experienced a 5–15% increase in wet days. During the MAM season, there was a 15–35% decrease in the number of wet days across South Sudan in the 1980s, and throughout Kenya, Uganda, Somalia, and southeastern Ethiopia in the 1990s. Again, during all four seasons in the 1980s, wet days decreased by 15–35% over Sudan and South Sudan. In the 1990s, the JJAS season experienced less than a 10% decrease in wet days in most of Ethiopia and the Lake Victoria basin. The DJF season experienced a 5–15% decrease in the number of wet days in Uganda over the 1980s, 1990s, and 2010s (Fig. 9p–t).

Figure 9
figure 9

Changes in seasonal mean state of number of wet days at 1 mm thresholds for the four decades (1981–1990, 1991–2000, 2001–2010 and 2011–2020), average of 40 years (1981–2020) and differences between current 20 year (2001–2020) and previous 20 years (1981–2000). The Brown color (negative values) indicates decreased wet days dry, while the green color (positive values) indicates increased wet days.

Figure 10 illustrates the observed rate of change (%) in the seasonal mean state of the number of dry spells. As seen in the wet day’s analysis already covered in Fig. 10 above, the dry spell results once more reveal considerable decadal (10 years) variability of dry spells over the IGAD region. Most areas of the region had a 5–20% decrease in the number of dry spells during the most recent ten years (2011–2020). The MAM and JJAS seasons have observed a 40% decrease in dry spells in recent years (2011–2020). Sudan, South Sudan, and northern Uganda observed a decrease of 35 per cent in dry periods throughout the MAM, JJAS, and OND seasons. During the DJF, the region observed less than 15 percent decreased dry spells, particularly in Kenya and Uganda (Fig. 10s). In the 1980s, a 10–35% increase in wet spells was observed over most of Uganda, South Sudan, and western Ethiopia. During MAM, dry spells decreased by 5–10% in most parts of Kenya and Sudan, eastern Ethiopia, and southern Somalia. The 1980s observed a broad 15–35% decrease in dry spells throughout South Sudan, a rain belt in southern Sudan, and western Eritrea. Dry spells increased by 5–10% in South Sudan, western Ethiopia in the 1980s (Fig. 10a), southeastern Ethiopia, and throughout Uganda, Kenya, and Somalia in the 1990s (Fig. 10b). In addition, coastal and northeastern Kenya, southern and central Somalia, and most of Ethiopia experienced less than a 15% increase in the 2000s.

Figure 10
figure 10

Changes in Seasonal mean state of Number of dry spells at 1 mm thresholds for the four decades (1981–1990, 1991–2000, 2001–2010 and 2011–2020), average of 40 years (1981–2020) and differences between current 20 year (2001–2020) and previous 20 years (1981–2000). The Brown color (positive values) indicates decreased dry spells, while green color (negative values) indicates increased dry spells.

Implication of wet days and dry spells on drought and floods patterns

Figure 11 presents the Inter-annual variability of wet days’ anomaly over five potential agricultural and food baskets in the region. The wet days anomalies were compared with the SPI signal to assess the implication of wet days on drought/flood patterns over El Gadaref in Sudan, El renk in South Sudan, Arsi in Ethiopia, Trans Nzoia in Kenya and Arua in Uganda. In general, the statistics reveal that the majority of the IGAD region experienced exceptional(unusual) drought/floods coincided with exceptional(unusual) decreased/increased wet days anomalies. The yearly variation in wet days anomalies is high across five sub-locations. These results confirmed the devastation drought occurred in the 1980s and exceptional wet conditions in recent years (2011–2020). In other words, the patterns of increase/decrease in wet days and dry spell anomalies significantly contributed to extra-ordinary drought and floods events in the past. For instance, MAM 2018 recorded extra-ordinary(unusual) increased number of wet days over Trans Nzoia in Kenya. The OND season in 1997 was the wettest year in the last 41 over most parts of the region, especially over Arua in northern Uganda and the highlands of western Kenya. This is in line with the conclusions reached by Ayugi et al.68, who found that the wettest years were 1997 and 1998, whereas drought occurrences were observed in the years 1987, 2000, 2006, and 2009 for SPEI-3 and 2000 and 2006 for SPEI-1268. The JJAS 2018 over Arsi in central Ethiopia recorded the highest wet days anomalies since 1981. The wet days and SPI anomalies over Upper Nile in South Sudan and Alqadaref in Sudan behind 1988 widespread floods. The increase in wet days in recent decades (2011–2020) was driven by extra-ordinary increased wet days in 2018, 2019 and 2020, while decreased wet days anomalies in the 1980s were driven by dry conditions in 1983–1986 over most parts of the region. Similarly, the severity of drought events in 1983, 1984, 1985 over Sudan, South Sudan, Ethiopia, 2011 Somalia, and 2021 over most parts of the region accelerated by a decreased number of wet days and prolonged dry spells. The decreased wet days anomalies in 1990/1991 over Alqadaref in Sudan exacerbated the extreme drought conditions in the record. These findings are consistent with those of Zhang et al.69, who determined that the most severe drought occurred in the 1980s, particularly in 1984, when more than 4.5 million people in Sudan were seriously afflicted and needed aid. The decreased wet days anomalies from OND 2020 to OND 2022 triggerred five consecutive failed rainy seasons over most parts of the IGAD region.

Figure 11
figure 11

Inter-annual anomalies of wet days over 5 potential agricultural sub-regions (the wet days computed over entire pixels in El Gadaref state in Sudan, Upper Nile County in South Sudan, Arsi zone in Ethiopia, Trans Nzoia county in Kenya and Arua district in Uganda) reference to 1981–2010 bassline.

Projected changes in wet days and dry spells patterns

Figure 12 presents the projected changes in wet days based on an ensemble of 10 of the best performed CMIP6 models over the IGAD region. The three future scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5) were analyzed for MAM (first row), JJAS (second row), OND (third row) and Annual (ANN) in the fourth row. The projected changes were carried out for the near future (2021–2050) and the far future or end century (2071–2100) relative to the 1985–2014 bassline(control) period. Except for the pattern of SSP2-4.5 for the far future), the future changes in wet days show a 10–20% decrease in wet days over Sudan, Eritrea, Djibouti and South Sudan under all three scenarios during the MAM season. Most parts of Kenya, Somalia, and southeastern Ethiopia, show an increased number of wet days. These findings varied from Ayugi et al.70 findings of a decrease in consecutive wet days (CWD) towards the end of the twenty-first century (2081–2100) relative to the baseline period (1995–2014) for MAM and OND over Kenya and Uganda. The discrepancy could be due to the models chosen, the baseline period, and the reference datasets. Similarly, our findings are consistent with Wainwright et al.51 findings of increased wet season rainfall. The JJAS season is projected to have a 30–40% increase in wet days over central and northern Sudan, a 10–20% decrease over South Sudan, the highlands of western Ethiopia and northern Uganda for both the near and far future. The 5–20% increase in wet days signal dominated the OND season over most parts of the IGAD region. On an annual time-scale, an increase in wet days is projected over Sudan, Eritrea, Djibouti, Somalia and coastal parts of Kenya, while a 20–40% decrease is projected over South Sudan, Uganda under SSP1-2.6 scenario and South Sudan under the SSP5-8.5. On an annual and across three seasons, Sudan is projected to have more than 50% increase in wet days under SSP2-4.5 scenarios.

Figure 12
figure 12

Spatial pattern of the projected number of wet days over the IGAD region is based on the CMIP6 ensemble mean for the near future (2021–2050) and far future (2071–2100) relative to the baseline period (1985–2014). The three future scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5) analyzed for MAM (first row), JJAS (second row), OND (third row) and Annual (ANN) in the fourth row.

Spatial pattern of projected number of dry spells over the IGAD region presented in Fig. 13. Under all scenarios, projected changes in dry spells show a 10–20% decrease over central and northern Sudan, most parts of Uganda, Kenya, Somalia and Djibouti, while 10–20% increase over southern parts of Sudan, most parts of South Sudan, highlands of western Ethiopia during MAM season. The changes in dry spells over Sudan seem to suggest shrinking in the dry season which extends from October to June. In Extreme northern Sudan, ASALs in Kenya, and southern Somalia are projected to have 10–20% increase in dry spells as opposed to a study by Ayugi et al.70, while a decrease is projected over southern and central parts of Sudan and most parts of Ethiopia across all scenarios during JJAS. These patterns suggest the prolonged dry season of JJAS over ASALs in Kenya and southern Somalia. The eastern Sudan, consistently projected to have increased dry spells, decreased over most parts of Uganda, Kenya, Somalia, southeastern Ethiopia during OND. Also, JJAS and OND patterns suggest the expansion of the rainy season over the Red Sea coastal parts in Sudan and Eritrea. Annually, dry spells are projected to increase under SSP1-2.6, decrease under SSP2-4.5 and SSP5-8.5 over most parts of the IGAD region. It appears that a rise in the number of rainy days and a decrease in dry spells explains the enhanced rainfall signals found by Mbigi et al.71 over Uganda and Kenya, Ogega et al.72 over Lake Victoria basin as well as the growing trend reported by Alaminie et al.73 over Ethiopia’s Upper Blue Nile Basin and Ngoma et al. over Uganda74.

Figure 13
figure 13

Spatial pattern of projected number of dry spells over the IGAD region based on CMIP6 ensemble mean for the near future (2021–2050) and far future (2071–2100) relative to the baseline period (1985–2014). The three future scenarios (SSP1-2.6, SSP2-4.5, SSP5-8.5) analyzed for MAM (first row), JJAS (second row), OND (third row) and Annual (ANN) the fourth row.

Source

Check new:   Poland's Position in FIFA Rankings: A Disappointing Sight for Sore Eyes