Viability of directly sown paddy rice

Methods

This study will look at the stand, yield and crop quality of rice (variety Hayayuki) under three different sowing methods and two water treatment regimes.
The three sowing methods are:
1) Standard method: Transplanted, 1 plant / 12 inches. Follow guidelines of “Rice growing manual of the Northeast” and our three years of experience.
2) Broadcast method: Broadcast seeded, 80 lbs/acre.
3) Row seeded method: Seeded in rows with a drum seeder, at 12 lbs/acre.

Water will be added gradually to all the treatment for the first four weeks. Then each treatment will all undergo two different water management regimes. These are:
1) Saturation: Pulse flooded in heavy rains, but wet for the majority of the time.
2) Flooded: Depth will range from 3-12 inches depending on the growing state. No water will be added in August as the canopy closes.

There are nine paddies on the farm. Three will be used for this experiment. These paddies have almost identical soil types and histories. They are all Buxton silt loam soil, with 10-12% clay content.
These paddies were finished in the spring of 2015 and received a combination of cover crops, flood fallowing and pig rotations. The pigs killed most perennial grasses. The paddies receive water from a pond uphill. This pond is fed from a spring, an ephemeral stream or our water pump, in severely dry periods. Each paddy has its own inflow and outflow. We will use six adjacent paddies in this experiment, with ten-foot berms separating them. Each is 5000 square feet with a height differential of no more than two inches within them. They were intentionally built and managed for experimentation and research.

Each paddy will be considered a block, three will be flooded and three saturated. The treatments will be arranged in a split-plot design, with water management as the main plot and seeding method as the sub-plot. Each treatment will be repeated three times, for a total of 18 subplots. We will be using a “Randomized Complete Block Design” and use ANOVA (alpha = 0.05) to analyze the data. Please reference Appendix A for a diagram. Each plot will be a 7 x 50 foot bed. The shape of the blocks is dictated by the ability to use the drum seeder, along with the ease of access for data collection, weeding and observation. There will be a three-foot aisle between each plot. To offset edge effects there will be a three-foot buffer from the edge of the paddy to the adjacent plot as well. A berm, with built-in drainage, will be created to enclose the treatments in each the flooded paddy. This will lessen the total area needed for flooding from 15,000 to 6,000 square feet and ensure adequate water for this experiment and the rest of our rice operation.

Measurements

Labor by activity will be tracked daily for any task related to the rice. All purchases for the rice project will be done electronically. Space in the high tunnel will also be calculated for opportunity costs.
The following data will be taken from two, half square meter quadrates in each treatment. Data will not taken from the outer rows of any treatment, to account for edge effect. This data will be pooled by subplot.
Heading dates – Collected when three panicles have emerged.
Tillering count: At heading date.
Maturity Date 1: When 50% of the rice is fully brown and has folded over.
Maturity Date 2: When 90% of the rice is fully brown and ready to thresh.
Lodging: Plants that have fallen between maturity dates.
Height: Recorded at Maturity Date 1.
Stand count: Recorded June 15, before we switch to the flooded and saturated water management regimes. Weed Data: Our baseline at planting/sowing will be no weeds. Weed counts will be taken every two weeks from late May till the end of July. The five most numerous weeds in our rice paddies (identified in 2015) will be tallied separately, and the rest tallied as “other”. After counts, hand weeding will occur in the row seeding and transplanting treatments. Hand weeding should correlate with weed pressure. Labor will be tracked for the hand weeding in the whole treatment. To emulate the reality of broadcasting at a normal scale, no weeding will take place in the broadcasting treatments. To obtain more data and help compare broadcast treatments, the above ground biomass will be taken from the weeds in each quadrate, dried and weighed during weed counts. Hand weeding will cease in the other two treatments around August 1. As the plant switches into a ripening life stage weed presence should have little effects on yields. At Maturation Date 1 the above ground biomass will be taken, dried and weighed in all treatments. It should be noted that we have found considerably less weed pressure and diversity in our paddies as compared to dry-land agriculture.

The Cono-weeder will be used in half of the non-experimental parts of our rice paddies and hand weeding in the other half. These labor times, recorded but unbudgeted, will allow us to anecdotally extrapolate weed labor with a conoweeder as compared to hand weeding.
Yields: Dry weight at 15% moisture content. The plants will be threshed by our foot-powered thresher and dried in our high tunnel.
Bobolinks caused major damage last season to our crop from August 25th to September 10th. Their damage was not uniform, as they prefer unripe seed. Six 50 x 50 foot nets will be put up to during time. No data collection will be taken at this time. Nets will be removed for data collection.

We will use new random quadrates for each data grouping. Each weed count, stand count and yield will get new quadrates. Heading dates and tillering will share a quadrate as will plant height, lodging and maturity dates.