Empirical Models of Seasonal Changes in the NDVI Vegetative Index in Spring Barley Crops

The study was conducted at the experimental base of the Menkovsky branch of the Agrophysical Research Institute in the Leningrad Region in 2015–2021. The purpose of the experiment is to develop a model of the seasonal dynamics of the NDVI index of spring barley crops. Comparing the current NDVI measurements during the growing season with the model data will allow monitoring and, if necessary, adjusting the condition of crops using agrotechnological techniques. Measurements were carried out once a week at permanent accounting sites in two field experiments. The NDVI dynamics for experiments with the cultivation of spring barley mixed with perennial grasses and in pure form were different. In the first case, the maximum value of the index occurred on the 28th week from the beginning of the year, in the second – on the 25th week. The indicated difference was due to the influence of perennial grasses, due to the increase in the aboveground mass of which the effects of herbicide treatment were leveled. Models of seasonal dynamics of NDVI, built on average long-term data, turned out to be more reliable for general analysis than one-year ones, since they reduced the influence of random factors. In both cases of modeling the variability of the seasonal course of the index, as applied to two experiments with the cultivation of spring barley, good results were obtained using a polynomial function of the 2nd degree (R² = 0,97 and 0,80, p = 0,00003 and 0,008). At the same time, complication of the model to a polynomial of the 3rd degree led to an increase in the accuracy (R² = 0,92) and reliability of the forecast (p = 0,004) only for the second experiment. In the models describing the seasonal dynamics of NDVI in individual years, higher accuracy corresponded to a polynomial of the 2nd degree.

spring barley (Hordeum vulgare L.), NDVI index, seasonal dynamics, empirical models, polynomial function