Cassava is an important staple crop for the rural population but also more and more used as a cash and industrial crop, contributing to economic development of sub-Saharan Africa (SSA). Therefore, sustainable intensification of cassava has become necessary, but the crop's yield potential is unknown. This study aimed to understand the yield potential of cassava and the dynamics of nutrient limitations in relation to water availability, to obtain better insight and a theoretical understanding of how the crop responds to nutrient availability and application of fertilizers in different agro-ecologies. Field experiments were conducted in different agro-ecologies: in Cross River (Tropical Rainforest), Edo (Forest Transition Savanna) and Benue (Guinea Savanna) in 2016 and 2017. Treatments included NPK fertilizers in four application rates, N, P and K omission plots and plots combining the highest rates of N and P (300 kg N ha-1 , 100 kg P ha-1) with respectively 60, 120, 180, 240 and 300 kg K ha-1 , and highest rate of N, P and K combined with secondary and micronutrients. Fresh and dry biomass and nutrient contents of leaves, stems and storage roots were recorded at 4 and 8 months after planting and at final harvest. At set times, light interception and detailed plant characteristics were measured, including canopy dimensions, number of leaves and leaf scars. The measured values were used to calculate other variables, including leaf angle distribution, light extinction coefficient, leaf area index (LAI) and radiation use efficiency (RUE). The LINTUL-Cassava model was recalibrated and tested for accuracy of water limited yield predictions. Further, model components were added that allowed us to compute N, P and K limited growth. Foremost, we evaluated the cassava yield potential and nutrient use efficiency when supplied with various N, P and K rates and at high N and P rates with increasing rates of K. Measured storage root dry matter (DM) yields ranged between 11 and 35 t DM ha-1. The largest yields were obtained at the largest rates of K with a mean agronomic efficiency of 60, 162 and 51 kg DM of storage roots per kg of N, P and K applied. Increasing the supply of K improved yields and therefore also the agronomic efficiency of N and P. Storage root yield responses to applied N, P and K fertilizers (2-18, 3-16 and 3-22 t DM ha-1 , respectively) varied across the agro-ecologies, reflecting variability in potential yields and applied NPK ratios. Addition of secondary and micronutrients did not affect cassava storage root yield. We found that the caloric energy yield of cassava per kg of N applied is 2.7 times larger than the value reported for maize. For the nutrient uptake dynamics in cassava, on average, 67, 61 and 52 % of total N, P and K uptake were already taken up at four months after planting (MAP), respectively. The maximum uptake rates for N, P and K were 0.21, 0.03 and 0.12 g m-2 d-1 respectively at about four MAP. Nutrient dilution curves for N, P and K at different application rates were quantified for the ...