Peanut production is vital to the Georgia agricultural industry, both economically and as a rotational crop for traditional row-crop producers. Georgia accounts for 49% of the peanut production in the U.S., with planted acreages averaging 231,000+ ha annually (2000–2009) (Georgia Peanut Commission, 2010). The majority (>90%) of peanut production in Georgia occurs in the Coastal Plain region. The highly-weathered, sandy soils and relatively high rainfall in this region are generally favorable for the crop. However, these soils, traditionally cropped under conventional tillage (CT) practices, tend to be drought-prone, and are susceptible to runoff, sediment, and chemical losses. Rainfall in the Coastal Plain (∼1250 mm/yr) tends to have short duration-high intensity, runoff producing storms especially at critical times during the growing season when antecedent water content is high and/or soil cover (residue, canopy) is none to minimal. Also, climatic conditions can yield extended drought periods during the crop growing season, thus, peanut production generally includes supplemental irrigation to minimize yield limiting water stress.

Conservation tillage (strip tillage, ST) adoption in the Coastal Plain region of Georgia has increased because ST systems accumulate residue and organic carbon at the soil surface, which dissipates raindrop impact energy and increases soil resistance; thus maintaining infiltration and reducing runoff, soil detachment, and sediment transport (Reeves, 1997; Truman et al., 2005; Truman et al., 2007; Truman and Nuti, 2010; Truman et al., 2011). However, studies have reported less runoff from CT systems than ST systems, especially 1–3 yrs after ST adoption due to increased consolidation (NeSmith et al., 1987; Radcliffe et al., 1988; Soileau et al., 1994; Cassel and Wagger, 1996). Consequently, paratilling, a non-inversion, deep tillage technique, is often used in ST systems to disrupt dense, water-restrictive subsurface layers. After fall paratilling in 2007, bulk densities for ST were as much as 31% less than those for ST prior to paratilling at the study site described herein. Disrupting these layers with paratilling reduces bulk density, increases infiltration, and reduces runoff (Truman et al., 2005; Truman et al., 2007). In the Coastal Plain region of Georgia, it is generally recommended that land managed in ST and planted to peanut be paratilled in the fall prior to planting peanut the following spring.

Frequency of severe rainfall events has increased throughout the U.S., including the Southeast, mainly in the form of increased intensity of extreme rainfall events (Nearing et al., 2005; Todd et al., 2006). Changes in rainfall intensity within a storm affect how rainfall is partitioned (infiltration, runoff) and sediment and chemical transport (Potter et al., 2010; Truman et al., 2007; Truman et al., 2011).

In the Coastal Plain region of Georgia, data are limited on effects of peanut cropping systems on runoff and sediment yields. Truman & Williams (2001) used simulated rainfall (I = 63.5 mm/h) to evaluate runoff and soil loss from single- and twin-row peanut systems as a function of canopy cover conditions. They found that single-row planted peanut had as much as 3-times more runoff and soil loss than twin-row planted peanut. Twin-row planted peanut maximized peanut canopy development and percent soil cover early in the growing season and minimized the time in which bare soil is vulnerable to a runoff producing rainstorm. However, they did not evaluate peanut production under strip-tillage nor with variable rainfall intensity patterns that commonly occur throughout the growing season. Our objective was to quantify runoff and sediment losses from a Tifton loamy sand managed under conventional- (CT) and strip- (ST) tillage systems and planted to peanut at planting, 30 days after planting, and after harvest.

The 1-ha field site was located near Tifton, GA (N 31° 26′, W 83° 35′) on a Tifton loamy sand (Typic Kandiudult; 82% sand, 7% clay; 2% slope), which represents over 762,000 farmable ha in the Coastal Plain region of Georgia. The site has been managed under CT and ST systems in a cotton (Gossypium hirsutum)-peanut (Arachis hypogea) rotation since 1998 (Potter et al., 2010; Truman et al., 2011). The CT consisted of fall disking, drilling a winter rye (Secale cerale) cover crop, followed by spring disking, turning with a moldboard plow, field cultivator levelling/conditioning, and bedding. Rye surface cover was incorporated 10–15 cm. The ST consisted of planting a winter rye cover just after crop harvest in the fall and killing the rye chemically 30 days before planting the next year's row crop. All ST plots were paratilled (depth = ∼30 cm) in the fall (11 October, 2007) prior to the 2008 peanut growing season. With ST, a 10-cm wide strip was tilled and used to plant the crop into. In 2008, this site was cropped to single-row peanut (planted 12 May, 0.9 m row spacing). Surface residue cover for CT and ST were <1 and 50%, respectively.

Rainfall simulation plots (2-m wide, 3-m long) were established on each treatment combination at three critical times during the 2008 peanut growing season: 13–20 May (freshly tilled seedbed condition at planting); 23–26 June (first fungicide application with minimal canopy cover); and 21–31 October (tilled/disturbed soil surface immediately after harvest with peanut hay distributed back over bed). Each 6-m² plot consisted of a wheel track and two half beds on either side of the wheel track. Antecedent water content (w) was determined gravimetrically (Gardner, 1986) from samples taken in three locations around each 6-m² plot just prior to simulating rainfall (0–2 and 2–15 cm depths).

Simulated rainfall was applied with an oscillating nozzle rainfall simulator (Truman et al., 2011) that used 80150 Veejet nozzles (2.3-mm median drop size). The simulator was placed 3 m above each 6-m² plot. Simulated rainfall was applied at three variable rainfall intensity (I_v) patterns: a spring time variable intensity pattern used at planting (I_vSPR); a summer time variable intensity pattern used at the first fungicide application (I_vSUM); and a fall time variable intensity pattern used just after peanut harvest (I_vFALL) (Fig. 1). Data from the I_vSPR pattern have been reported previously (Potter et al., 2010).

Variable rainfall intensity patterns were developed from measured 5- and 1-min natural rainfall data (30 yrs) collected at Tifton, GA. Natural rainfall during the spring (March, April, May), summer (June, July, August), and fall (September, October, November) were analyzed to determine respective patterns that occurred most frequently during each 3-month season. Parameters (I_max, time to I_max, Precip_max, duration) were averaged for the group of natural storms occurring most often during each 3-month season. For each season, the individual storm with the most parameters similar to the average of the entire group of storms was selected and its pattern programmed into the simulator on a 1-min basis as the representative I_v pattern for that 3-month season (I_vSPR for spring; I_vSUM for summer; I_vFALL for fall).

Rainfall duration for the I_vSPR, I_vSUM, and I_vFALL patterns was 70, 70, and 60 min, respectively. Average rainfall I during the I_vSPR, I_vSUM, and I_vFALL patterns was 56, 30, and 36 mm/h, respectively. Maximum rainfall I during the I_vSPR, I_vSUM, and I_vFALL patterns was 160, 102, and 132 mm/h, respectively. Total rainfall volume applied to each 6-m² plot was repeatable (avg. I_vSPR = 61 mm, CV = 3%; avg. I_vSUM = 30 mm, CV = 3%; avg. I_vFALL = 36 mm, CV = 2%). Irrigation well water was used in all simulations. Runoff (R) and sediment (E) were measured from the down slope end of each 6-m² plot at 5-min intervals throughout each simulation, and determined gravimetrically. Infiltration (INF) was calculated by difference (rainfall-runoff). The parameter dINF was calculated by difference (maximum INF–minimum INF). Estimates of days of water for crop use for I_vsum events were calculated from INF values for each tillage treatment and an assumed evapotranspiration value (6 mm/day).

Treatments consisted of two tillage systems (CT, ST) and three rainfall intensity patterns (I_vSPR, I_vSUM, I_vFALL). Means, coefficients of variation (CV, %), and standard error bars are given for measured data (n = 3). Unpaired t-tests (two-tailed distribution) were used to determine significance among treatment means using SigmaStat 3.1 (Systat, 2004). All test statistics were evaluated at P = 0.05, unless otherwise noted. All other data analysis was conducted with Microsoft Office Excel 2007.

We quantified runoff and sediment losses from a Tifton loamy sand managed under CT and ST systems and planted to peanut at three times during the peanut growing season (at planting, 30 days after planting, after harvest). Rainfall simulation events with variable rainfall intensities (I_v) representative of each time period (spring, I_vSPR; summer, I_vSUM; fall, I_vFALL) were used. We found that over the three seasonal events, runoff and infiltration from CT and ST plots was 9–22% and 78–91% of the total rainfall applied, respectively. The most runoff occurred from CT-I_vFALL plots; the least from ST-I_vSUM plots. Sediment yields from CT and ST plots ranged from 105–1419 kg/ha. The most sediment occurred from CT-I_vSPR plots; the least from ST-I_vSUM plots. On a per unit rainfall applied basis, the most sediment occurred from CT-I_vFALL plots. Thus, CT-I_vFALL plots yield the most runoff and sediment per unit of rainfall applied among treatment combinations evaluated. Runoff and sediment curves for CT and ST plots had similar shapes as corresponding simulated rainfall patterns. Rainfall intensity and runoff from CT and ST plots for I_vSPR, I_vSUM, and I_vFALL events were correlated (r = 0.73–0.95). Runoff and sediment yield from CT and ST plots for I_vSPR, I_vSUM, and I_vFALL events were also correlated (r = 0.66–0.99). Values of R_max for CT and ST plots were 14–18% of the maximum intensity (160 mm/h) for I_vSPR events, 12–20% of the maximum intensity (102 mm/h) for I_vSUM events, and 7–19% of maximum intensity (132 mm/h) for I_vFALL events. Values of R_max and E_max for CT and ST plots occurred 5–8 min after the maximum intensity for I_vSPR events (20 min); 3–4 min after the maximum intensity for I_vSUM events (47 min); and 6–8 min after I_vFALL events (27 min).

Strip-tillage reduced runoff for I_vSUM, and I_vFALL events, and sediment losses for all three events. Over all events, CT plots averaged 60% more runoff, 86% higher R_max values, 3.3-fold more sediment, and 3.7-fold higher E_max values than ST plots. Maximum difference between CT and ST plots for runoff (2.4-fold), R_max (2.7-fold), and sediment (3.8-fold) was for fall events; maximum difference between CT and ST plots for E_max (5.6-fold) was for summer events.

Frequently occurring rainfall intensity patterns that occur during critical times (planting, first fungicide application ∼30 days after planting, immediately after harvest) during the peanut growing season influences infiltration, runoff, and sediment yields. Our results show the pronounced effect of the three rainfall intensity patterns and tillage on runoff and sediment losses from those periods during the peanut growing season. Understanding the magnitude of runoff and sediment losses help define “windows of risk” and how conservation tillage can reduce those losses and help producers manage natural resources and agrichemical losses from peanut cropping systems.