Predicting ratoon rice growth rhythmbased on NDVI at key growth stages of main rice

RESEARCH Predicting ratoon rice growth rhythm based on NDVI at key growth stages of main rice Kai-lou Liu1, Ya-zhen Li1*, Hui-wen Hu1 The growth of ratoon rice (Oryza sativa L.) is affected by growth of main rice, so it could be evaluated by determining growing condition of main rice. Objective of the study was to find regression model for predicting ratoon rice growth rhythm through establishing the quantitative relationship between growth rhythm in ratoon rice and normalized difference vegetation index (NDVI) after heading of main rice. NDVI at key growth stages of main rice was measured using active spectrometry for 12 varieties in 2008 and for 23 varieties in 2009, and compared the physiological indexes of main and ratoon rice. There was an exponential correlation (P < 0.05) between the NDVI after heading of the main rice and biomass, N concentration of green leaves and stems or leaf area index (LAI) after heading of the ratoon rice for all the 12 varieties in 2008. The regression models from the rice variety experiment in 2008 were used to calculate predicted values with NDVI after heading of main rice in 2009. The results showed that the predicted values of biomass, N uptake, and LAI in ratoon rice were significantly different to measured values. However, there were exponential and significantly positive correlation (p < 0.05) between biomass and N uptake and LAI in ratoon rice and NDVI after heading of main rice in 2009. Therefore, it suggested that regression models were not perfect, and need to be improved adding more parameters about spectral characteristic of main and ratoon rice. Key words: Canopy NDVI, exponential correlation, ratoon rice, regression equation. INTRODUCTION Ratoon rice (Oryza sativa L.) is a rice cropping system that has twice harvests in one season as it develops from dormant buds on the stubbles after main rice is harvested. It has been practiced in China since the West Jin dynasty (AD 265-316) (Guo, 1993) and latterly planted in India, Thailand, USA, and the Philippines (Nakano and Morita 2007). It is called twice harvesting rice in Japan (Nakano et al., 2008). The grain yield of ratoon rice is usually 1500 to 3000 kg ha-1, being much lower than main rice. As ratoon rice has higher nutrient using efficiency per unit time and per unit land area than normal rice, many new varieties have been developed with high grain yield of ratoon rice and it is reported that yields of ratoon rice is as high as 8100 kg ha-1 in Youxi county, Fujian Province of China (Xu et al., 2015; Lin et al., 2015). In general, leaf area index (LAI) of ratoon rice was about one-eighth to one-fifth of that of main rice, and grain number per panicle and yield of ratoon rice were about one-third of those of main rice, while photosynthetic rate National Engineering and Technology Research Center for Red Soil Improvement, Jiangxi Institute of Red Soil, Nanchang 330046, China. *Corresponding author (). Received: 16 October 2014. Accepted: 29 June 2015. doi:10.4067/S0718-58392015000500005 1 410 of ratoon rice was 20% to 70% higher than that of main rice in 10 d after full heading and net assimilation rate of ratoon rice was 3.7 to 6.0 times that of main rice from booting to maturity (Yi et al., 2009b). Many studies have been conducted to explore the reasons for the high yield of ratoon rice (Jiang et al., 2003; Zhang et al., 2005; Chen et al., 2008; Yi et al., 2009a). The grain yield of ratoon rice was affected by varieties, sowing time, cultivation method, fertilizer rate, and stubble height of main rice. Yi et al. (2009b) suggested varieties with more grains would achieve higher yield of ratoon rice. Zhang et al. (2005) found that the grain yield was higher when main rice was sowed earlier. Chen et al. (2008) indicated that the grain yield under super high-yield cultivation relied less on the photosynthate in the period of grain filling, which was beneficial to steady and high yield of ratoon rice. Jiang et al. (2003) suggested that the grain yield of ratoon rice had parabola correlation with N application for bud development. Xu et al. (2008) found appropriate amount of N application for bud development would be needed when ratoon rice was to be cultivated with varieties having spikelets per panicle of main rice. Harrell et al. (2009) indicate that when the initial stubble height is reduced from 40 to 20 cm the growth of the ratoon crop is altered by shifting panicle point of origin during the early growth period and delaying maturity. Yi et al. (2009a) suggested decline of the stubble height resulted in decreased number of effective panicles and increased CHILEAN JOURNAL OF AGRICULTURAL RESEARCH 75(4) OCTOBER-DECEMBER 2015 grain number per panicle in ratoon rice. When ratoon rice is used as forage rice, to produce the highest total DM yield, Nakano and Morita (2008) suggest that the best time for first harvest is the full heading stage and the best N management is to provide 300 kg N ha-1 in accordance with 33.3 N application at transplanting, 33.3% at the maximum tiller number stage, 16.7% at 7 d after the first harvest, and 16.7% at 21 d after the first harvest, respectively. And, the total estimated leaf and stem total digestible nutrient yields were higher in double harvesting than in single harvesting in most cultivars in Japan. Spectral determination provides an automatic, quick, and nondestructive method of assessing crop biomass and nutrient status. The normalized difference vegetation index (NDVI) is an empirically derived index used to estimate plant biomass through the integration of the redvisible and near-infrared spectral regions to represent plant pigmentation and chlorophyll content respectively. NDVI is one of the most extensively applied vegetation indices related to leaf area index (LAI), biomass and predict yield (Raun et al., 2001; Ma et al., 2001; Inman et al., 2007; Gnyp et al., 2014). Wang et al. (2003) and Guo et al. (2008) found that there were significantly positive correlation between chemical contents and N nutrition of leaves and NDVI in rice and corn. Li et al. (2008) suggested that there was exponential function relationship between N absorption and NDVI in wheat. NDVI measurement has been applied to diagnose nutrient, particularly N status for the growth period development of many crops such as rice and wheat (Zhu et al., 2008), but it is never reported for predicting ratoon rice growth from main rice growth. Axillary buds and nutrient in stubble were important for growth and high yield of ratoon rice, because axillary buds in ratoon rice were affected by total content of internodes after heading of main rice. Total N content of internodes was the main factor to decide ratoon rate of axillary buds in early productive stages in main rice, but soluble sugar content in middle and late productive stages in main rice (Yi et al., 2005). Zheng et al. (2004) suggested that the development of axillary buds depends on the remnant root syste (...truncated)