Abstract:There is a variation in onset dates between the two definitions across regions; with meteorological and crop‐germination definitions of onset showing differences. The standard deviation of onset dates shows East‐Central‐South gradient. CMIP6 model's representation of trend in onset differs strongly from reanalysis data, with significant intermodel differences. Prediction of seasonal onset is crucial to agriculture in southern and eastern Africa. Here, we applied two definitions of onset, namely meteorological and agricultural (crop‐germination), to evaluate CMIP6 models through the lens of rainfall onset over representative maize agricultural regions of South Africa, Tanzania, Malawi and Zambia. We use the ERA5 reanalysis as a proxy for observations, and robust regression to calculate a statistical comparison of the onset definitions for the period 1979–2021. Evaluation of ERA5 reanalysis shows similar magnitude and pattern as gauge based MSWEP. Our results show that, for meteorological onset, Johannesburg, with a subtropical highland climate, experienced earliest onset after 23 December; and an increasing trend (later onset) but not statistically significant (p = 0.2). Over Bethlehem, which has continental climate, the earliest onset date was after October 9 and an increasing interannual variability since 2000 is noted. The standard deviation of onset dates across the regions shows an East‐Central‐South gradient. We also found that the crop‐germination onset definition shows earlier onset of seasonal rains, it differs considerably across regions, and has higher interannual variability, in comparison with the meteorological definition. Over Lilongwe, Mbeya and Lusaka, late meteorological onset with a weak positive and insignificant trend is observed. The CMIP6 model's representation of onset trend differs from reanalysis data, with inter‐model differences. Late meteorological onset is underestimated by GFDL‐CM4 and MPI while INM5, MPI and NorESM overestimate the observed earliest onset. The largest bias is shown by INM and MPI which simulate earliest and latest onset as 190 (07 January) and 206 (23 January) respectively. In addition, models often fail to simulate sufficient precipitation to produce onset for seed germination and crop development. The ACCESS model showed an insignificant trend (p value = 2) and later onset over Lilongwe, an insignificant trend (p value = 0.9) over Lusaka, and an earlier onset over Mbeya. Using the agricultural onset definition, over Bethlehem, all the models and the ERA5 reanalysis did not produce enough precipitation to meet onset conditions. We suggest that rainfall onset studies use several definitions or metrics of onset and that the choice of metric be informed by the research question. Using such an ensemble of onset metrics contributes to a better understanding of variability and uncertainties in agricultural productivity.

Predicting Rainy Season Onset in the Ethiopian Highlands for Agricultural Planning

Integrating climate prediction and regionalization into an agro-economic model to guide agricultural planning

Variability of rainy season onsets over East Africa

Seasonal prediction of Horn of Africa long rains using machine learning: The pitfalls of preselecting correlated predictors

Applications of machine learning to predict seasonal precipitation for East Africa

Probabilistic rainy season onset prediction over the greater horn of africa based on long-range multi-model ensemble forecasts

Assessment of Seasonal Rainfall Prediction in Ethiopia: Evaluating a Dynamic Recurrent Neural Network to Downscale ECMWF-SEAS5 Rainfall

Assessing the Climatological Characteristics of Observed and Simulated Seasonal Onset of Precipitation Over Southern and Eastern Africa

Locally defined seasonal rainfall characteristics within the Horn of Africa drylands from rain gauge observations

Markov chain analysis of dry, wet weeks and statistical analysis of weekly rainfall for agricultural planning at Dhera, Central Rift Valley Region of Ethiopia

Artificial intelligence models for prediction of monthly rainfall without climatic data for meteorological stations in Ethiopia

Beyond the regional average: Drivers of geographical rainfall variability during East Africa's short rains

Discovering causal factors of drought in Ethiopia

Modeling the Spatio-Temporal Meteorological Drought Characteristics Using the Standardized Precipitation Index (SPI) in Raya and Its Environs, Northern Ethiopia

Rainfall variability and its seasonal events with associated risks for rainfed crop production in Southwest Ethiopia

Regionalization of the Onset and Offset of the Rainy Season in Senegal Using Kohonen Self-Organizing Maps

Analysis of the long-term agricultural drought onset, cessation, duration, frequency, severity and spatial extent using Vegetation Health Index (VHI) in Raya and its environs, Northern Ethiopia

Seasonal precipitation forecasting for water management in the Kosi Basin, India using large-scale climate predictors

Forecasting vegetation condition for drought early warning systems in pastoral communities in Kenya

Seasonal variability and predictability of monsoon precipitation in Southern Africa

GIS-based spatio-temporal analysis of rainfall trends under climate change in different agro-ecological zones of Wolaita zone, south Ethiopia