Search

• Knowledge of onset and cessation of rains is critical for crop planning in tropical regions. • Analyses of these parameters at production-relevant scale are lacking. • We quantify interannual variability and trends at agroecosystem level. • Results indicate that existing agricultural practices may need to be revised. • Extended weather and sub-seasonal forecast services should be implemented in the area. Analysis of Rainfall onset date (OD), cessation date (CD), and length of growing period (LGP) for specific sites in highly dissected topography and highly variable climate may not provide actionable information for crop production planning. In tropical highland regions information on these parameters is scant at a resolution relevant for targeted management. This study examined recent (1981-2016) OD, CD, and LGP variability and trends for the main rainy season in different agroecosystems (AESs) in the northwestern Ethiopian Highlands. Onset criteria were derived from surveys, rainfall data, and previous literature whereas cessation criteria were set from the soil water holding capacity (WHC), daily reference evapotranspiration (ETo), and daily rainfall in each site. Dry spells (DS) were analyzed for the small rainy season in higher elevation AESs where the season is relevant for potato production. All analyses were performed using site specific data grouped by agroecosystem (AES), a unit that has similar climate, soil, crop, and farm management for better agricultural decisions. Results show high inter-annual variability of OD and CD, and LGP exhibited a significant trend in some AES and greater variability in higher elevation AES. Generally, trend analysis results showed early onset and cessation of rainfall. Significant increasing trends and variabilities in DS and OD may significantly affect crop production and thus AES specific crop production calendar should be revised to minimize crop failure. The analysis also confirmed that farmers perception is consistent with meteorological analysis. The results emphasize the importance of AES-based improved seasonal weather forecasts and tailored climate information services to guide farm decisions and improve management of climate variability by smallholder farmers. It also concluded that AES level analysis can better provide actionable information for decision makers and growers than site specific and scattered studies as mosaic results are reported between sites grouped in the same agroecosystem. [ABSTRACT FROM AUTHOR]