DynACof process-based model parameterization and validation for robusta coffee

Coffee Arabica and Robusta are the main species for coffee production. Both these species are grown in different suitable climate. For instance, suitable environmental conditions for Robusta species are found in lowland areas of Vietnam with relatively high temperatures and precipitation. Although, this species is resilient to climate change, but the impact of climate change has been documented in the literature. Moreover, until now, models have applied water management as an adaptation approach. However, no model has considered Robusta coffee cultivation under agroforestry systems. Therefore, the present study was conducted to modify the DynACof model to simulate the Robusta coffee in agroforestry systems. It is important to mention that dataset used to modify this model was obtained from Robusta coffee open sun system due to lack of information on agroforestry systems. The study was conducted in ten districts of Vietnam where climate is highly suitable for Robusta coffee cultivation. The DynACof model was previously developed for Arabica coffee thus our aim was to modify the model to improve its capabilities for Robusta coffee. The climate data for historical period was retrieved from European Centre for Medium-Range Weather Forecasts, Reanalysis v5 (ERA5). The statistical indicators were applied to evaluate the model performance based on observed and predicted bean yield. In most districts, the model achieved a high coefficient of determination (R²), approximately 0.5, along with low error in some districts, ranging from 8.7% to 21.5%, in districts such as Dak Song, Krong Nang, Lam Ha, Duc Trong, and Krong Buk. The D-Index indicated the reliable predictions, with most values falling in the range of 0.7 to 0.8, particularly in districts like Dak Song, Dak Glang, Krong Nang, and Chu Prong. Some districts, such as Lagari, Duc Trong, and Lam Ha, examined the high nRMSE values ranging from 29.6% to 35.5%. Thus, the residual plot shows the downward trend in error distribution. On average, the model predicted a bean yield of 2,700 kg/ha in districts where the bean maturity duration was around 260 days with an optimal climatic condition. These results demonstrate the potential of the model for evaluating the coffee yield and guiding sustainable management strategies to address the challenges posed by climate change.

Coffee Arabica and Robusta are the main species for coffee production. Both these species are grown in different suitable climate. For instance, suitable environmental conditions for Robusta species are found in lowland areas of Vietnam with relatively high temperatures and precipitation. Although, this species is resilient to climate change, but the impact of climate change has been documented in the literature. Moreover, until now, models have applied water management as adaptation approach. However, no model has considered Robusta coffee cultivation under agroforestry systems. Therefore, the present study was conducted to modify the DynACof model to simulate the Robusta coffee in agroforestry systems. It is important to mention that dataset used to modify this model was obtained from Robusta coffee open sun system due to lack of information on agroforestry systems. The study was conducted in ten districts of Vietnam where the climate is highly suitable for Robusta coffee cultivation. The DynACof model was previously developed for Arabica coffee thus, our aim was to modify the model to improve its capabilities for Robusta coffee. The climate data for historical period was retrieved from European Centre for Medium-Range Weather Forecasts, Reanalysis v5 (ERA5). The statistical indicators were applied to evaluate the model performance based on observed and predicted bean yield. In most districts, the model achieved a high coefficient of determination (R²), approximately 0.5, along with low error in some districts, ranging from 8.7% to 21.5%, in districts such as Dak Song, Krong Nang, Lam Ha, Duc Trong, and Krong Buk. The D-Index indicated the reliable predictions, with most values falling in the range of 0.7 to 0.8, particularly in districts like Dak Song, Dak Glang, Krong Nang, and Chu Prong. Some districts, such as Lagari, Duc Trong, and Lam Ha, examined the high nRMSE values ranging from 29.6% to 35.5%. Thus, residual plot shows the downward trend in error distribution. On average, the model predicted a bean yield of 2,700 kg/ha in districts where the bean maturity duration was around 260 days with an optimal climatic condition. These results demonstrate the potential of the model for evaluating the coffee yield and guiding sustainable management strategies to address the challenges posed by climate change.