V-Mango model: an intricate 3D simulation of the King of Fruits

Mangoes (Mangifera indica) are known as the King of Fruits, a title which suits them well. Mango trees can live up to 300 years in tropical and subtropical countries and there are over 400-500 varieties of mangoes. It is not only a tasty tropical fruit but a cup of mangoes contains half of the daily C vitamin and 7% of daily fiber intake requirement. Some of the main challenges of mango production are pest damage and inconsistent fruit production.

Boudon and colleagues, based in France, recently published the V-Mango model which intricately simulates mango tree growth in 3D. The model enables mango growers to estimate fruit production and predict dates when the plant is susceptible to pest damage. The leading scientist, Dr Frédéric Boudon works on plant modelling and computer science at CIRAD and co-authored the open-access software, OpenAlea for plant growth modelling.

Mango tree growth cycles can be divided into four developmental phases: vegetative growth (9 months), resting period (2 months), flowering period (2-3 months) and fruit growth (4 months). One cycle lasts to 1.5 years and the first two phases can overlap on the same tree.

The researchers built a computer model using Functional-Structural Plant Model (FSPM) approach where the plant development was made up of growth units (GU). Scientists previously monitored how mango trees grow on the Reunion Island, east of Madagascar. These measurements were used by Boudon and colleagues for modelling.

Increasingly bushy computerised mango trees — The architecture of mango trees simulated with V-Mango model during the four developmental stages (initial structure, flush growth during vegetative growth, flowering period and fruit harvest period). Source: Boudon et al., 2020.

The model captures intricate details of mango tree development daily and considers the effect of temperature on tree growth, burst date, full bloom, harvest date and yield. The growth of different plant parts (i.e. leaves, branches, flowers) are simulated by sub-models relying on light capture, allometric equations, carbon and water-related biophysical processes. The fruit weight is calculated from cell division and expansion and also stimulates accumulation of organic compounds (sucrose, fructose) and minerals (K, Mg, Ca). After one growth cycle (1.5 years) the model then considers the previous year’s growth to predict next year’s.

Comparison of observations with simulations — The V-Mango model simulates mango tree growth daily in 3D. The development phases range from the appearance of the main axis until the mature growth unit. Source: top photographs from Dambreville et al., 2015 and model pictures from Boudon et al., 2020

The model was used to simulate two scenarios. Firstly, an orchard of 100 mango trees and could clearly identify on which days the tree’s growth is most impacted by mango blossom gall midge, Procontarinia mangiferae. Secondly, the model was used to estimate how large the fruiting branches need to be for optimal fruit growth. Based on 1,000 simulations, the trees produced between 300-400 fruits per cycle.

The scientists explain that this study “constitutes a basis for the development of agronomic tools to design cultivation practices aiming at maximising and making more regular mango yield, and reducing pesticide use”. However, the authors also add, “to make this model a useful tool for agronomy, the effects of cultivation practices such as pruning should be integrated, as well as interactions with pests”.

Models are only as good as the data and assumptions made for different processes. This study shows how fundamental research on plant growth can be used for high-quality modelling which can help fruit tree growers. View a related poster presentation here as a summary of their approaches.