Temporal distribution of rice sheath mite Steneotarsonemus spinki Smiley ( Acari : Tarsonemidae ) as influenced by the climatic parameters in the lowcountry wet zone of Sri Lanka

The rice sheath mite (RSM; Steneotarsonemus spinki Smiley) is considered as the most destructive mite pest of rice (Oryza sativa L) in most of the rice-growing countries including Sri Lanka. Under heavy infestation of RSM, yield losses could reach up to 95%. Intermittent heavy showers negatively affects the RSM population while prolong dry periods influence positively on its population growth. However, no detailed studies have been done in Sri Lanka to elucidate the influence of weather conditions on population dynamics. This study was conducted in three different locations in the low country wet zone (LCWZ; elevation < 300m amsl; annual rainfall > 2,200mm) from 2013-2015 to examine the relationship between selected weather parameters and population dynamics of RSM. An experiment was laid out in a randomized complete block design in three replicates where crop was planted in 15-day intervals under pesticidefree condition. In weekly intervals, 25 leaf sheaths from 10 randomly selected plants from each plot were observed under microscope to count the number of mites. Climatic data were collected from respective meteorological units for the study period. The results revealed that, rainfall (r=-0.543) during the 2013/2014 Maha season at Bombuwala and RH (r=-0.86) during the 2013 Yala season at Labuduwa have a negative correlation with the RSM population, while temperature between 29 – 32 °C (r=0.794 and 0.902 at Bombuwala and Labuduwa, respectively, during the 2013 Yala season) and more sunshine hours with dry weather conditions affected favourably. Abundance of RSM-infested host plants in the previous season had a significant influence on its occurrence in new rice crop under favourable environment conditions.

The first published reference to S. spinki as a pest of rice crop was from southern China in 1968 and is considered a serious pest of rice in tropical Asia, Central America and the Caribbean (Hummel et al., 2009).In Sri Lanka, RSM was first recorded in 2000 Sri Lanka Journal of Food and Agriculture (SLJFA) ISSN: 2424-6913 Journal homepage: www.slcarp.lk in the intermediate zone (Nugaliyadda et al., 2005).Subsequently, it has become the most important and destructive mite pest in most of the ricegrowing areas in the island.
The RSM lives in space between rice leaf-sheaths, but can also live under the hull of rice grains and survives in stubbles and ratoon during the off season.The RSM feeds on cells of leaf sheaths, stems and kernels at the reproductive stage of rice plant leading to impaired grain development with empty or partially-filled grains with brown spots and panicle standing erect.Among the mites, S. spinki (family Tarsonemidae) infests flag leaf sheath causing brown discoloration.Its damage at early reproductive stage of the crop develops chocolate-brown coloured lesions on the leaf sheath and the feeding on the floral parts causes considerable grain sterility (Mutthuraju et al., 2014).It also damages the rice crop indirectly by vectoring and facilitating the establishment of pathogens such as sheath rot caused by S. oryzae and panicle blight causing Burkholderia glumae (Hummel et al., 2009).Yield losses in rice due to RSM damage has been estimated as 5 -95% in a number of countries including Sri Lanka (Navia et al., 2010;Pushpakumari et al., 2010).Hummel et al. (2009) reported that Japonica varieties in India are more susceptible to RSM than the Indica varieties, with a 20% yield loss.
Facultative pathogenesis is reported in S. spinki.All the descendants of virgin females are males.The mother female then mates with its male offspring's to produce both male and female mites.A mated female mite on an average could produce 55 eggs in the laboratory at 24.5 -35 °C.The mites multiply during the whole year, with a minimum of 6 generations per month in spring and less than 3 in winter (Hummel et al., 2009) while Increasing temperatures results in shorter generation times.Hence, the RSM is able to produce 48-55 generations per year under ideal climatic conditions.Three potential pathways of RSM entry have been identified i.e. via rice seeds (Rao et al., 2000), natural means such as transfer by wind, water, insects and birds (Almaguel et al., 2003) and mechanical means such as farming equipment (Pushpakumari et al., 2010).
Studies on host range of S. spinki have been conducted only in few countries yet.American wild rice (O.latifolia Desv) is an alternate host for S. spinki in Costa Rica and Panama (Sanabria and Aguilar, 2005) while Cynodon dactylon (L.) Pers.
Application of chemical control measures after the heading stage of crop growth is not worth as the grain damage has already occurred.Identification of S. spinki damage at the early infestation stage is difficult and hence, most of the farmers tend to use different pesticides even after soft dough stage of the crop.Therefore, identification of its population dynamics is of significance as it help farmers taking prompt and correct decisions to manage the pest.This study was conducted in LCWZ for 40 consecutive months starting from December, 2012 to determine the temporal distribution of S. spinki in relation to climatic factors.

Experimental sites:
The studies were conducted at three research stations in the low country wet zone (LCWZ) i.e.

Experimental design and crop sampling:
Experiments were laid out in a randomized complete block design replicated three times using the rice (Oryza sativa) variety Bg300 (three-months age class).Plot size was 5 m × 4 m.The crop was established at 15-day intervals by transplanting 18 to 21-days old seedlings.Experiment was repeated three times since 2013 where, every time after 3 rd month of the study, the experimental field had six rice crops at different maturity levels.The trials were maintained under rainfed condition and was free of pesticide use.Fertilizer application and other agronomic practices were done according to the recommendations of the Department of Agriculture, Sri Lanka.Weed management was done manually.
Samples were collected at weekly intervals starting from 4 weeks after planting (WAP) and continued till harvesting.As the crop was established in 2week intervals, different crop growth stages were found in the experimental field at each sampling time.Hence, all the plots existed at a given sampling time were considered for sampling, and the average value of a parameter was considered for analysis.Sample size was 10 tillers (approximately 30 sheaths) per plot.Then, 25 sheaths were taken from each sample to record the number of RSM and its eggs, which were observed under microscope.

Collection of weather data:
Weather data for the whole study period was collected from the meteorological unit situated at each experimental site.Weekly total rainfall and weekly averages of maximum temperatures, rainfall, relative humidity, sunshine hours and wind velocity were recorded as meteorological data,.
Data analysis: Pearson correlation analysis was carried out for whole study period and for each cropping season separately, to identify the correlation between the moving average of RSM and weather parameters.Then, multiple regression was carried out along with the testing of multicollinearity of predictor variables for the total evaluation period to find out the most contributing factors for the moving average of RSM.Furthermore, stepwise multiple regression was carried out for the whole study period and also for separate seasons by using SAS 9.1.3. to identify the most contributing factors for RSM counts.

Results and Discussion
During the whole study period at Bombuwela and Bentota, there were number of PRM population peaks indicating its population build up with the time (Figures 1 and 2).The results suggested that the RSM population in the previously planted crops in the experimental field would have influenced the population density in the subsequent crops.However, at the Labuduwa site, high population densities of S. spinki were seen only at the beginning of the experimental period.A low RSM population was experienced also in farmer fields during the year 2014 and 2015 in the region where Labuduwa site is located.

Effect of weather parameters on rice sheath mite population:
The results of multiple regression and multicollinearity analysis for the data from Bombuwela location showed that tolerance for the predictor variable ranged between 0.5 to 0.9 and variance inflation parameter ranged between 1.1 -1.9.This indicates that there is no considerable multicollinearity in the predictor variables.Data on S. spinki at Labuduwa also indicated that there were no considerable multicollinearity as predictor variables resulting in 0.88 -0.71 tolerance and variance inflation parameter was ranged in between 1.12 -1.39.The stepwise multiple regression carried out for the whole evaluation period and also for each season separately.The results of stepwise multiple regression carried out for both locations and the Pearson correlation between S. spinki population and weather parameters are shown in Tables 1 and 2, respectively.Relative humidity: A negative contribution of relative humidity on S. spinki population was observed during the whole study period (Table 1).The relative humidity (RH) was the only factor which was included to the regression model (Y= -4.37 RH + 482.54,R 2 =0.326,P=0.024).However, there was no any predictor variable selected for regression equation with Labuduwa data for the whole study period.Pearson correlation with S. spinki population and relative humidity also indicated a negative correlation among the said factors (Table 2).
During the whole study period at Bombuwela and the 2013 Yala season at Labuduwa, a significantly negative correlation was observed between the relative humidity and mite population (Figure 3).In the other seasons, the correlation was not statistically significant.This could be due to the effect of other weather parameters.Further, S. spinki could survive in a wide range of relative humidity levels even from 77% to >92% depending on the availability of other weather parameters at favourable levels and the availability of the host crop (Figure 4).Thus, no conclusions could be drawn regarding the critical minimum humidity level they can survive as the RH values recorded during the whole study period were higher than 77%.-1270.24,R 2 =0.8136,P=0.001) indicated that there is a positive contribution of the average temperature to S. spinki population (Table 1).However, the contribution of temperature on RSM population at Bombuwela during the 2014 Yala season showed a negative contribution (Y = -14.98Temp.+ 503.05,R 2 =0.346,P=0.01).
Correlation results presented in Table 2 also indicated the same trend.Therefore, temperature could positively contribute to RSM population only until a certain level.The temperature variation prevailed at each location in the whole study period was low (28.6 °C -32.7 °C; Figure 4).The results revealed that high RSM populations could be observed within the temperature range of 30 °C -32 °C.However, future studies under a wider temperature variation are needed to determine the critical temperature for RSM.

Rainfall:
The RSM population is widely known to have a negative correlation with rainfall.Like most of the delicate soft-bodied insects, sheath mites can also be flushed off by heavy rains other than the mechanical damage.Furthermore, heavy rains   There was a significantly negative correlation among rainfall and RSM population during the 2013/2014 Maha season at Bombuwela (Table 2).Though there were occasions that showed negative correlation, such relationships were statistically not significant.This may be due to the influence of another parameter or due to unknown factors, which need to be studied further.A higher density of S. spinki was observed under rainy condition indicating the ability of the mites to adapt to a wider range of weather conditions.

Wind velocity:
According to the stepwise multiple regression analysis, wind velocity had a negative contribution to RSM population (Y = -25.02WV + 261.39,R 2 =0.294,P=0.048) during the 2015 Yala season at Bombuwela while a positive contribution (Y = 154.54WV + 10.82, R 2 =0.37,P=0.027) during the 2015/2016 Maha season at the same location (Table 1).Pearson correlation analysis also showed a significantly negative correlation of wind velocity during 2015 Yala season and positive correlation during 2015/2016 Maha season at Bombuwela (Table 2).Therefore, there could be a critical range of wind velocity for the population build-up of RSM.
The S. spinki populations were highly concentrated within the wind velocity range of 1.5 -3.5 km/h (Figure 5), providing some evidence on the favourable range of wind velocity for the mite species.However, further studies are needed to provide a firm conclusion on this.The RSM can be transferred by wind (Rao et al., 2000) and thus, there is a possibility of RSM moving away from the existing host under high wind velocity.This could be the reason for comparatively low population densities recorded under high wind velocity (> 3.5 km/h) than that of under moderate velocities (1.5-3.5 km/h).

Figure 4 .
Figure 4. Relation between relative humidity and ambient temperature on RSM population at Bombuwela.
the egg development of sheath mites resulting in low population density.

Figure 5
confirmed this by showing more S. spinki population towards periods with lower rainfall.

Figure 5 .Figure 6 .
Figure 5. Relation between rainfall and wind velocity on RSM population at Bombuwela.Sun shine: Fluctuation of S. spinki population in relation to the number of sunshine hours per day during the whole study period at Bombuwela is illustrated in Figure6and Table1.The results revealed high S.

Table 1 .
Regression equation fit by stepwise multiple regression for weather parameters and RSM populations at Bombuwela and Labuduwa

Table 2 .
Pearson correlation between RSM population and weather parameters