Using growing degree days for predicting Itchgrass (Rotboellia cochinchinensis) emergence in the tropics

Abstract

Models based on thermal time describing seedling emergence are valuable not only to understand important aspects of the biology of a weed species, but also for decision making when implementing weed control practices. Most seedling emergence models have been developed in temperate regions where extreme temperatures during the winter and summer prevent seed germination and seedling growth of summer and winter weed species, respectively, creating relatively defined seedling emergence periods. Conversely, to the extent of our knowledge, there are no weed seedling emergence models for tropical conditions, and it is unknown how the absence of extreme temperatures throughout the year might affect weed seedling emergence patterns in agroecosystems. We studied seedling emergence of itchgrass (Rottboellia cochinchinensis), which is an aggressive grass weed in multiple agronomic crops such as corn, rice, cotton, and sugarcane in the tropics. Field experiments were conducted in several locations in Costa Rica from March to August and August to December in 2010 and 2011. Itchgrass seedling emergence timing and soil temperature were monitored after tillage. Most of the emergence occurred two weeks after tillage, regardless of the location and time of the year, suggesting that tillage promoted itchgrass germination. Gompertz and Weibull models were adjusted to the observed cumulative seedling emergence based on thermal time using growing-degree days (GDD) with a base temperature of 20 C, which was determined using an iteration procedure. The models satisfactorily described the observed emergence (r 2=0.79 to 0.93) although the Weibull model best fitted the data based on the Akaike information criterion (AIC). Fifty percent emergence (E50 ) occurred approximately at 115 GDD although there were differences between locations and time of the year, thus E50 ranged between 64 to 300 GDD. Additionally, two phases of rapid seedling emergence occurred consistently in all sites and years, suggesting the presence of two groups of itchgrass individuals with different seed germination requirements within the evaluated populations. The present study illustrates how predictive seedling emergence models can be generated using GDD for tropical conditions.