RESUMO
Cassava-maize intercropping is a common practice among smallholder farmers in Southern Nigeria. It provides food security and early access to income from the maize component. However, yields of both crops are commonly low in farmers' fields. Multi-locational trials were conducted in Southern Nigeria in 2016 and 2017 to investigate options to increase productivity and profitability through increased cassava and maize plant densities and fertilizer application. Trials with 4 and 6 treatments in 2016 and 2017, respectively were established on 126 farmers' fields over two seasons with a set of different designs, including combinations of two levels of crop density and three levels of fertilizer rates. The maize crop was tested at low density (LM) with 20,000 plants ha-1 versus high density (HM) with 40,000 plants ha-1. For cassava, low density (LC) had had 10,000 plants ha-1 versus the high density (HC) with 12,500 plants ha-1.; The fertilizer application followed a regime favouring either the maize crop (FM: 90 kg N, 20 kg P and 37 kg K ha-1) or the cassava crop (FC: 75 kg N, 20 kg P and 90 kg K ha-1), next to control without fertilizer application (F0). Higher maize density (HM) increased marketable maize cob yield by 14 % (3700 cobs ha-1) in the first cycle and by 8% (2100 cobs ha-1) in the second cycle, relative to the LM treatment. Across both cropping cycles, fertilizer application increased cob yield by 15 % (5000 cobs ha-1) and 19 % (6700 cobs ha-1) in the FC and FM regime, respectively. Cassava storage root yield increased by 16 % (4 Mg ha-1) due to increased cassava plant density, and by 14 % (4 Mg ha-1) due to fertilizer application (i.e., with both fertilizer regimes) but only in the first cropping cycle. In the second cycle, increased maize plant density (HM) reduced cassava storage root yield by 7% (1.5 Mg ha-1) relative to the LM treatment. However, the negative effect of high maize density on storage root yield was counteracted by fertilizer application. Fresh storage root yield increased by 8% (2 Mg ha-1) in both fertilizer regimes compared to the control without fertilizer application. Responses to fertilizer by cassava and maize varied between fields. Positive responses tended to decline with increasing yields in the control treatment. The average value-to-cost ratio (VCR) of fertilizer use for the FM regime was 3.6 and higher than for the FC regime (VCR = 1.6), resulting from higher maize yields when FM than when FC was applied. Revenue generated by maize constituted 84-91% of the total revenue of the cropping system. The highest profits were achieved with the FM regime when both cassava and maize were grown at high density. However, fertilizer application was not always advisable as 34 % of farmers did not realize a profit. For higher yields and profitability, fertilizer recommendations should be targeted to responsive fields based on soil fertility knowledge.
RESUMO
Cassava is an important crop in the developing world. The goal of this study was to review published cassava models (18) for their capability to simulate storage root biomass and to categorize them into static and dynamic models. The majority (14) are dynamic and capture within season growth dynamics. Most (13) of the dynamic models consider environmental factors such as temperature, solar radiation, soil water and nutrient restrictions. More than half (10) have been calibrated for a distinct genotype. Only one of the four static models includes environmental variables. While the static regression models are useful to estimate final yield, their application is limited to the locations or varieties used for their development unless recalibrated for distinct conditions. Dynamic models simulate growth process and provide estimates of yield over time with, in most cases, no fixed maturity date. The dynamic models that simulate the detailed development of nodal units tend to be less accurate in determining final yield compared to the simpler dynamic and statistic models. However, they can be more safely applied to novel environmental conditions that can be explored in silico. Deficiencies in the current models are highlighted including suggestions on how they can be improved. None of the current dynamic cassava models adequately simulates the starch content of fresh cassava roots with almost all models based on dry biomass simulations. Further studies are necessary to develop a new module for existing cassava models to simulate cassava quality.