EFFECTS OF AZOLLA BIOMASS GROWTH ON FLOOD WATER TEMPERATURE AND pH, TILLERING AND YIELD OF PADDY RICE [EFECTO DEL CRECIMIENTO DE LA BIOMASA DE AZOLLA EN LA TEMPERATURA DEL AGUA DE INUNDACIÓN Y pH, MACOLLAMIENTO Y RENDIMIENTO DE ARROZ]

Ministry of Agriculture, Livestock and Fisheries, P.O. Box 30028, Nairobi, email: woyange@yahoo.com,+254711748464; Department of Plant Science and Crop Protection, University of Nairobi, P.O. Box 30197, email: george.cheminingwa@uonbi.ac.ke, +254721723806; School of Biological Sciences, University of Nairobi, email: jiykaya@yahoo.com,+254722307387; Department of Agricultural Resource Management, University of Embu, P.O. BOX 6 60100. Embu, Kenya, Tel: +254-727 933950, +254720487126, Email, paulnjiruh@gmail.com, *Corresponding author


INTRODUCTION
Azolla is an aquatic fern mainly found in tropical and temperate paddies and forms a symbiotic association with nitrogen fixing blue-green algae, Anabaena azollae.In Mwea Irrigation Scheme paddies, Azolla biomass coverage ranges from 25% to 50% at peak times.It has a fast biomass doubling rate and this has caused a challenge in its management making the farmers to have a negative perception about it.The fast multiplication rate of Azolla makes it a noxious weed (Ivens, 1989).
According to Campbell (2011), it doubles its biomass within 7-10 days under high nutrient levels and temperatures above 18 °C.Kitoh et al. (1993) reported a doubling rate of 2-3 days under laboratory condition, using polluted pond water while Hussner (2010) reported a doubling rate of 3-10 days.In Mwea Irrigation Scheme, Kenya, the doubling rate was found to be 4-6 days under average temperatures of 24 °C, relative humidity of 75% and pH of 6.85.Azolla biomass growth rate is affected by the environmental conditions and increases with light intensity and temperature (Wagner, 1997).The fast multiplication rate can lead to coverage of as high as 100 %, with a 30 cm thick mat which can block light, prevent photosynthesis and seriously affect oxygen diffusion (Landsdown, 2015).The biomass coverage shields light and hinders heat uptake from the environment thus reducing flood water temperatures.Fosu-Mensah et al. (2015), reported a reduction in flood water temperatures by 1.9-2.0°C.
Studies have also shown that a full cover of Azolla significantly increases tiller count and grain yield but reduces flood water pH by 1.4 -2.2 (de Macale and Vlek, 2004).In Kenya, studies on Azolla in Ahero research station have been done by incorporating it in paddies.However, physio-chemical characteristics and rice growth have not been conducted.The objective of this study was to determine the effects of Azolla biomass growth on flood water temperature and pH, tillering, adventitious root development and yield of paddy rice.The results of this experiment are therefore expected to enlighten farmers on the increased nitrogen efficiency uptake, increased tillering and grain weight of paddy rice.

Experimental site
The study was conducted in Mwea Irrigation Scheme, at Mwea Irrigation and Agricultural Development Centre (MIAD), in Kirinyaga County located at an altitude of 1159 metres above sea level.Mwea lies within Agro-ecological zones LM3 and LM4 (marginal cotton zones).The climate is tropical with equatorial and medium high altitude characteristics and receives an average of 930 mm of rainfall out of which 510 mm is during the long rainy season, with 66 % reliability.The mean temperature is 22 o C with a minimum of 17 o C and a maximum of 28 o C. The soils are predominantly imperfectly drained vertisols (LB 8).

Experimental design and treatments
Treatments comprised Azolla inoculated and noninoculated plots (control), laid out in a random complete block design, with three replications.In the Azolla treated pots, 40 g of fresh Azolla biomass (adequate to attain a total surface coverage of 500g/0.9m 2 within 20 days) was inoculated in plastic pots, each measuring 30 cm x 30 cm, filled to 20 cm mark with top soil of Mwea rice growing area.Non treated control pots were not inoculated with Azolla biomass.Nitrogen was applied at the recommended Mwea levels (Wanjogu et al., 1992) of 60 kg N ha -1 in three splits: at transplanting, 21 days after transplanting and 55 days after transplanting.Phosphorus (P2O5) and potassium (K2O) were supplied also at recommended Mwea levels of 50 kg ha -1 , each as triple super phosphate and muriate of potash respectively to all plots at seedling transplanting time.Rice variety Basmati 370 obtained from MIAD Seed Production Unit, and Azolla collected from the drainage canal within MIAD fields, were used for the trial.Two rice seedlings each were transplanted per pot and weed control was done manually at 21 and 35 days after transplanting.Water levels in the pots were maintained at 2.5-5 cm above soil level throughout the growing season.
About 5 g of fresh Azolla biomass was also inoculated in 7 plastic pots flooded with 3 litres of canal water obtained from MIAD irrigation canal (replenished every 3 days) and left to grow for 10 days to determine the doubling time.Temperatures and relative humidity for the growth period were recorded and final Azolla fresh biomass weight taken on the 10 th day after inoculation.Azolla doubling time was determined based on the formula DT = t/r (Kitoh and Shiomi, 1991) where: DT is doubling time, t = duration of growth; r = log wt./w0 / 0.301; wt.= weight of Azolla at time t; w0 = weight of Azolla at zero time.

Data collection
Flood water temperatures were recorded at 9 am and 3 pm on 21, 35, 45 and 60 days after transplanting, by partial immersion of an ordinary mercury thermometer.Three measurements were taken from different points in each pot at each recording time.About 50 mls of flood water was drawn using a glass beaker on 21, 31, 45 and 60 days and pH determined using glass electrode method (Jackson, 1958).Ten hills per treatment were sampled to determine plant height and tiller numbers at 21, 30, 42, and 55 and 75 days after transplanting, corresponding to rooting, tillering, maximum tillering, panicle initiation and heading stages respectively.
Plant height was determined by measuring from ground level to the vertical tip of flag leaf, tillers numbers were obtained by physically counting tillers in each hill and averaging for 10 hills.Productive panicles were counted from 10 hills and number of spikelets in each panicle determined.Grain yield was determined by harvesting plants from 10 hills per treatment, threshing and weighing grains at adjusted moisture content of 14%.Unfilled grains were separated using a salt solution of specific gravity 1.01 and filled grains counted using an electronic multi-auto grain counter (Everwell corporation-Tokyo Japan) and the number of spikelets per panicle determined.Grain weight (g) was determined by weighing 1000 filled grains (using an electronic grain counter) and adjusting the weight to 14% moisture content.At 120 days after transplanting (maturity time), plants were excavated using a panga then uprooted by hand, root biomass separated from the shoot and both oven dried to constant weight at 100 0 C for 28 hrs (Faichney and White, 1983).Root and shoot dry weights were then obtained using a scale (Salter Bracknell digital scale).Basal node length was determined from 5 plants per hill, for 10 hills, using a measuring rule.The numbers of adventitious roots per node were obtained by physically counting the roots with the aid of a hand lens (Eschenbach, Germany).

Data analysis
Data collected were subjected to analysis of variance using Statistical Analysis System SAS (version 9.0) statistical application software.Post hoc analysis (where there was significance) was carried out using the least significant difference test at p≤0.05.

Daily temperatures
The average temperatures for season 1 ranged between 21.6 and 24.0 0 C, increasing steadily during heading (shown by red arrow).During season II, the weekly average temperatures were higher but gradually decreasing at heading time.The average temperatures during rice crop growth stage for the first season were lower than the second season.At reproductive stage, average temperatures for the second season were however, higher than in the first season.

Azolla biomass doubling rate
The doubling rate of Azolla biomass ranged from 4.0 to 5.7 days in the first season and 4.4 to 5.9 days in the second season (Table 2).

Linear regression relationship between doubling time and average temperatures
The correlation coefficient between doubling time and temperature during the first season (r=0.18) was stronger than in the second season (r=0.07) as shown in Fig 1 .2aand 1.2b.

Effect of Azolla biomass on flood water temperature and temperature fluctuations
During the first season, flood water temperature readings for Azolla inoculated samples recorded a pvalue of 0.021, 0.030 and 0.035 at 21, 32 and 42 DATs respectively, which were significantly lower than that of the control (Table 3).In the second season, there was also significant reduction in flood temperatures in Azolla inoculated pots at 32 and 60 DATs (Table 3).The daily temperature fluctuations were significantly higher for the Azolla inoculated pots than non inoculated pots at all sampling points in the first season and at 32 and 60 DAT in the second season.In Azolla inocluated pots, flood water temperatures fluctuated from 1.5 to 5.6 o C and 1.1 to 3.9 o C for the first and second seasons respectively (Table 4).

Effect of Azolla biomass cover on flood water pH
Flood water pH was significantly lower in Azolla inoculated pots than the controls at all sampling points, in the two seasons of research.The pH differences between Azolla inoculated pots and the control ranged from 0.7 to 1.7 in the first season and 0.6 to 0.9 in the second season (Table 5).

Effect of Azolla growth on adventitious roots of rice plants
During both seasons, plants in Azolla inoculated pots showed significantly more adventitious roots per node, more and longer roots/mm 2 than non-inoculated control plants.Azolla inoculation increased adventitious roots per node and roots/m 2 , compared to non-inoculated control (Table 6 and Figure 2).In Azolla inoculated pots, rice plants basal internodes were significantly shorter than non-inoculated plants.

Effect of Azolla biomass on root length, root biomass and plant biomass
During the two seasons under consideration, roots were significantly shorter in Azolla inoculated pots than in non-inoculated control pots (Table 7).Azolla inoculation had a significantly shorter root length than in the control.Also, Azolla inoculation significantly increased root biomass and plant biomass compared to non-inoculated control.Tiller numbers and plant height for the Azolla inoculated pots also significantly increased.

Effect of Azolla on yield components and grain yield
The effect of Azolla inoculation on panicle length, spikelets/panicle, number of panicles, grain weight, percentage grain filling and yield, is shown in Table 8.Azolla inoculation, with a p-value of 0.140 and 0.52 had no significant effect on panicle length during season one and two respectively.Number of spikelets per panicle was significantly higher in the Azolla y=0.22x -0.06, r 2 =0.032, p=0.41 y=0.142 x+0.14x, r 2 =0.0053, p=0.35 inoculated pots than in non-inoculated control pots, during the second season.However, there were no significant differences during the first season between the Azolla inoculated and non-inoculated control plots in this parameter.Azolla inoculated pots showed higher number of panicle/m 2 and 1000-grain weight than non-inoculated control during the first season.
The second season showed no significant differences with respect to the number of panicles/m 2 and 1000grain weight.Although there were no significant differences in percentage grain filling and grain yields during both seasons, grain yields varied from 4.4-5.65 t/ha and 2.3 to 2.8 t/ha during the first and second season respectively.

DISCUSSION
Azolla double its biomass within 5.5 to 6.7 days.The rate was faster during the first season than the second.The doubling rate was within the range of 3-10 days reported by Hussner (2010).The relatively faster doubling rate during the first season than in the second season can be attributed to higher temperatures, which may have enhanced enzymatic, photosynthetic and stomatal conductance thus increasing growth rate (Sage and Kubien, 2007).Azolla has been reported to tolerate a wider range of temperatures but with an optimum of between 18-27 °C (theazollafoundation.org).Temperatures average of 22±0.21 °C in Mwea Irrigation scheme is therefore conducive for Azolla biomass accumulation, which may be beneficial for increased soil nutrient status.However, findings in Mwea Irrigation scheme in 2016, showed a reducing trend of biomass levels attributed to water shortage.Azolla biomass growth reduced flood water temperatures and increased daily water temperature fluctuations by an average of 1.1 ± 0.31 °C and 2 ± 0.24 °C respectively.The reduction in flood water temperature can be attributed to the presence of Azolla coverage on flood water surface which shielded flood water from solar radiation.Temperatures in the water therefore remained lower than that of the control.The evapo-transpiration effect of the plant biomass, which draws heat energy for evaporation from the water body, also increased cooling of the water body.The temperature reduction effect was higher in the first than in the second season.This may be attributed to higher average temperatures and lower relative humidity in the first season than in the second, which enhanced evapotranspiration and thus increased the flood water temperature.Fosu-Mensah et al. ( 2015) reported a reduction in flood water temperatures of between 1.9 and 2.0 °C under Azolla cover.The author attributed this to prevention of rapid heating of the flood water by Azolla canopy.Azolla cover reduced flood water pH by between 0.7 and 1.7.Reduction in pH can be attributed to the effect of Azolla canopy, which reduced flood water temperature thereby minimizing ammonia volatilization from the water body.Reduced ammonia loss increased flood water N content and had the associated effect of lowering flood water pH.The reduction in flood water pH due to Azolla biomass cover which limits surface N losses has been reported by Fosu-Mensah et al. (2015).
Azolla growth enhanced the number of adventitious roots in rice plant and reduced the basal internode length.Increase in number, length of adventitious roots and shortening of basal stem internode from Azolla treated plots can be attributed to the effect of reduced oxygen balance resulting from Azolla biomass growth in flood water.Landsdown (2015) showed that Azolla cover on flood water hinders oxygen diffusion.This may have stimulated a physiological complementation strategy of development of shorter basal stem internode for growth of adventitious roots to access oxygen.This adaptive mechanism towards augmentation of depleted oxygen supply has been reported in cereals by Atkinson (2014).Rice plant therefore can physiologically adapt to depleted oxygen supply through the development of adventitious roots.
Azolla growth reduced fibrous root length and enhanced plant biomass.The significant reduction in root length can be attributed to Azolla biomass canopy effect, which caused a reduction in water temperatures.This is hypothesized to have reduced enzymatic activities slowing physiological processes of growth and resulting into less root length development.The effect of lower temperature causing a reduction in root length has been reported by Essemine et al. (2010).Increase in plant biomass can be attributed to increased N efficiency uptake due to reduced loss of N as a result of low pH as reported by De Macale and Vlek (2004).The increased N levels may have also been supplemented by direct excretion of nitrogen from the Azolla anabaena system.Wagner (1997) reported that Azolla anabaena system directly excretes about 5% of its nitrogen fixed into the water body.This enhanced vegetative growth and resulted into increased biomass for the Azolla treated plots).
Azolla biomass enhanced growth and yield components of rice crop.During the first season, Azolla treatment significantly increased plant height, panicle/m 2 and individual grain weight.In the second season, tiller numbers and spikelets/panicle were increased by Azolla treatment.The enhancement of growth and yield components may be attributed to the effect of Azolla moderating the flood water temperature and hindering pH increase.The reduced temperature and lowered pH may have had the effect of reducing ammonia volatilization and consequently increasing efficiency of nitrogen utilization from applied fertilizers and N excreted from Azolla anabaena symbiosis.De Macale and Vlek (2004) reported increased efficiency of N utilization resulting from reduced pH.Wagner (1997) also reported that up to 3-4% N from the Azolla anabaena symbiosis can be directly excreted into the water body to the benefit of the rice crop.Azolla growth in flood water is therefore beneficial to farmers through increased fertilizer efficiency uptake, increased grain yield and biomass, which is becoming more useful in straw livestock production.

CONCLUSION
Azolla biomass growth in flooded paddies shields direct sunlight, consequently reducing flood water temperature, with a net effect of reducing N loss from water body.This enhances efficiency of N uptake and increases paddy rice yields Fig 1.2a: Linear relationship between doubling time and average temperatures during the first season

Figure 2 .
Figure 2. Effect of Azolla biomass growth on basal node and adventitious root growth

Table 1 .
Average temperatures of irrigation water and relative humidity during 2017 at MIAD centre Heading time

Table 2 .
Azolla biomass doubling rate at MIAD centre, Mwea, during the year 2016

Table 3 .
Effect of Azolla biomass on flood water temperatures (°C) in Mwea, during the year 2016

Table 4 .
Effect of Azolla biomass on flood water temperatures fluctuation ( 0 C) in Mwea during the year 2016

Table 5 .
Effect of Azolla on flood water pH in Mwea during the year 2016.

Table 6 :
Effect of Azolla on adventitious roots and basal internode length of rice plants in Mwea during the year 2016.

Table 7 .
Effects of Azolla on rice plant root length, root biomass, plant biomass, tiller numbers and plant height in

Table 8 .
Effect of Azolla biomass growth on yield components and grain yield in Mwea during the year 2016