Near-Surface Soil Nutrient Changes over Time under Native Prairie and Managed Agriculture in Arkansas

Landuse change from native prairie to managed agriculture can have substantial impacts on soil nutrient properties. Nutrient release from soil organic matter decomposition is the soil’s inherent source of long-term fertility; thus it is imperative to understand the effects of continued landuse over time to avoid mistaking actual soil property changes with simple inter-annual variability from one year to the next. The objective of this study was to evaluate the effects of landuse (i.e. managed agriculture and native prairie) in two contrasting physiographic regions (i.e. the Ozark Highlands region of northwest Arkansas and the Grand Prairie region of east-central Arkansas) on the change in extractable soil nutrients over a 15-yr period from 2001 to 2016. Extractable soil Ca, Mg, and Zn increased at least two times more over time (P < 0.05) under cultivated agriculture in the Grand Prairie than under native prairie in the Grand Prairie or either landuse in the Ozark Highlands. Averaged across landuse, extractable soil S increased nine times more over time (P < 0.05) in the Ozark Highlands than in the Grand Prairie, while extractable soil Na and Mn increased at least six times more over time (P < 0.05) in the Grand Prairie than in the Ozark Highlands. Averaged across region, extractable soil Mn increased 2.5 times more over time (P < 0.05) under native prairie than under agricultural landuse. Results from this long-term field study clearly demonstrate how landuse and regional soil characteristics can affect near-surface soil nutrient contents, which should be taken into consideration when implementing conservation and/or ecosystem restoration activities.


Introduction
Grasslands are one of the most abundant terrestrial ecosystems on Earth, and the most abundant ecosystem in North America [1]. The grasslands of the Great Plains region of North America once extended from Canada to the United States-Mexico border and from the base of the Rocky Mountains to Wisconsin.
Specifically, tallgrass prairies were one of the main North American grassland ecosystems, where tallgrass prairies were dominant in the eastern third of the Great Plains [1]. Naturally dense grasslands are widely biologically diverse in flora and macro-and micro-organisms and provide an array of ecosystem functions, such as animal habitat, resource production, soil and water conservation, recreation, and nutrient cycling [2]. Tallgrass prairies are often dominated by C4 grasses, such as big bluestem (Andropogon geradii), switchgrass (Panicum virgatum), and indiangrass (Sorghastrum nutans) [3].
The large and dense accumulation of organic matter associated with the fibrous roots of monocot grass species contributes to soil organic matter (SOM) and the overall stability of grassland soils, which in turn aid in the reduction of erosion and the regulation of soil water and gas exchange in an environmentally appropriate manner [2]. Grassland root systems affect soil structure through soil penetration as roots grow and proliferate, soil moisture alteration, organic compound exudation, and contribute large quantities of organic matter, carbon (C), and nutrients to the soil through root decomposition [4]. Annual senescence of aboveground vegetation in prairies often allows a thick layer of litter to accumulate on the soil surface that, in the absence of natural fire or a prescribed burn, slowly decomposes to provide another substantial source of organic material and nutrients to eventually be assimilated into the SOM pool and the soil in general.
Coupled with soil pH, which affects microbial activity, SOM decomposition, and cation exchange capacity (CEC), the SOM fraction is responsible for the supply and retention of essential plant nutrients, such as Ca, Mg, and K, and is the source of soil C. However, in regions of the once-prominent and expansive tallgrass prairie, where the climate is relative warm and humid, such as in east-central Arkansas, rapid SOM decomposition associated with landuse change has been shown to impact soil pH and the long-term storage of soil C and N [5], but can also negatively affect the long-term storage of other essential plant nutrients. Brye and Moreno [6] concluded that the resilience of the soil of a native tallgrass prairie is governed by time, where soil biological activity alteration may occur relatively quickly, while soil physical and chemical property changes occur much more slowly.
Brye et al. [7] evaluated soil quality and the relationships among selected soil properties across a climosequence of tallgrass prairies on loamy-textured soils in Arkansas. Results showed that soil quality differed between physiographic regions/climates (i.e. the cooler and drier Ozark Highlands region of northwest Arkansas and the warmer and wetter Grand Prairie region of east-central Arkansas) [7]. Soil organic matter, C, and extractable soil nutrients were generally were lower and soil pH, EC, and extractable soil P, K, Ca, Mg, and Fe were greater under tilled agriculture than native prairie landuse [8].
Brye and West [9] characterized the effects of conversion from native prairie to agricultural grassland management on soil surface properties in the Ozark Highlands and deduced that soil surface properties in agriculturally managed grasslands were similar to those in nearby native prairies. The conversion from native prairie to grazed and ungrazed forage landuse increased soil pH and extractable soil P, Mg, and Mn in the foragelands compared to adjacent native prairies [9]. In contrast, extractable soil S was greater in the native prairie than in the forageland, while soil EC, extractable soil K, Ca, Na, Fe, Zn, and Cu, and TN, TC, SOM, and C:N ratio were unaffected from native prairie conversion to forage landuse [9]. The removal of vegetation by haying, as examined by Brye and Moreno [6] in east-central Arkansas, can also affect soil quality in many ways. The long-term productivity of a grassland ecosystem is affected by the duration of annual surface vegetation removal, whether live or dead, and impacts the near-surface soil C balance and nutrient cycling [6].
More recently, McKee et al. [5] [5]. Soil organic matter was also 2.5 times greater in all prairie sites sampled in 2016 in both physiographic regions compared to under managed agriculture in Grand Prairie [5].
In addition, soil pH under managed agriculture was greater (pH = 6.7) than under native prairie (pH = 4.7) in the Grand Prairie [5].
Based on the landuse effects on the change in soil pH, SOM, C, and N over time, as reported in McKee et al. [5], it is likely that plant macro-and micronutrients were negatively affected as well. Given that the release of nutrients from SOM oxidation is the soil's inherent source of long-term fertility, hence sustainability, it is imperative to understand the effects of landuse change over sufficient time so as to not mistake apparent soil chemical property changes with simple inter-annual variability from one year to the next. Therefore, the objec- to 2016. It was hypothesized that pastureland and managed agricultural landuse will have greater extractable soil nutrient contents compared to native prairie landuse due to management practices and inputs. It was also hypothesized that landuse and region combined will have a more significant impact on soil nutrients compared to region and landuse alone.

Regional Characteristics
The Ozark Highlands (36˚N -38˚N lat., 91˚W -95˚W long.), major land resource area (MLRA) 116A, is located within southern Missouri, northeastern Oklahoma, and 23% of the MLRA is located within north-central Arkansas [10]. As a whole, MLRA 116A is approximately 2.1 million ha [10], where 93% of the land is privately owned, with 32% consisting of grassland [11]. The landscape of the Ozark Highlands, which is part of the Springfield Plateau, is variable. Steep, forested slopes descend into stony valleys and historic prairies and sedimentary rocks dominate, but dolostone, sandstone, limestone, and shale also comprise much of the underlying bedrock in the region [10]. Oak (Quercus spp.), hickory (Carya spp.), and shortleaf pine (Pinus echinata) make up the major tree species, while fescue (Lolium arundinaceum), an introduced species, now dominates many of the managed grasslands [10] [12]. Soils in MLRA 116A are typically weathered from limestone and/or a medium-to fine-textured cherty residuum, typically resulting in shallow to deep Udults and Udalfs [12]. Grand Prairie was once a much larger native tallgrass prairie spanning ~130,000 ha, but has subsequently been reduced to <1% of the original land area due to extensive conversion to row-crop production in the region [13]. The Grand Prairie is part of the Mississippi Alluvial Plain, where typically deep Udalfs are present [10] [12]. Hardwood vegetation, such as loblolly pine (Pinus taeda), shortleaf pine, cherry bark oak (Quercus pagodifolia), southern magnolia (Magnoila grandiflora), and cottonwood (Populus deltoides), is also native to and widely distributed in the Grand Prairie and throughout MLRA 134 in general [12].
The climate within both regions is humid temperate. The Grand Prairie is generally warmer than the Ozark Highlands, with mean annual air temperatures of 16.6˚C and 14.5˚C, respectively [14]. Similarly, the Grand Prairie is generally wetter than the Ozark Highlands, with mean annual precipitation of 126 and 116 cm, respectively [14].

Site Descriptions
Chesney and Stump Prairies (Table 1; Figure 1) are located in the Ozark Highlands region of Benton County, Arkansas, which are fragmented remnants of the historic Lindsley Prairie. In northwest Arkansas, the Lindsley Prairie once extended across more than 4000 ha [15]. The Chesney prairie specifically supports over 450 plant species, 29 of which are native and 18 plant species are recognized as rare [15]. Big bluestem, little bluestem (Schizachyrium scoparium), Indiangrass, switchgrass, large flower tickseed (Coreopsis grandiflora), prairie grayfeather (Liatris pycnosachya), and rattlesnake master (Eryngium yuccifloium) are some of the typical species present in the Chesney Prairie [16]. Sager Creek, an ephemeral stream, divides the Chesney Prairie and many prairie mounds are also present [17]. The soil of the Chesney and Stump Prairies has a loamy texture covering cherty limestone, allowing for moderately to well-drained conditions and moderate permeability [18]. Jay silt loam (fine-silty, mixed, active, thermic  Oxyaquic Fragiudalfs [18]) is mapped throughout both prairie sites in the upland site positions with a udic soil moisture regime. The approximate elevation for the Stump and Chesney Prairies is 362 m above sea level [9].
Immediately adjacent to the remnant Stump and Chesney Prairies is agriculturally managed land of varying disturbance levels, thus contrasting landuses exist within the same soil map unit (Table 1; Figure 1). A managed, ungrazed grassland and a managed, grazed pasture are located immediately adjacent to the Stump Prairie [9]. The managed grassland is predominantly tall fescue, where aboveground vegetation had been removed by haying multiple times a year for the last 20 years prior to soil sampling for this study [9]. The managed pasture has been used for approximately the last 20 years to graze a small head of cattle multiple times per year prior to soil sampling for this study [9]. However, the managed pasture had never been cultivated, while the managed grassland has only been cultivated a few times in the last 20 years for replanting perennial grasses. At Chesney Prairie, which is approximately 2 km north and slightly east of the Stump Prairie, an area of previously cultivated agriculture exists within the Chesney Prairie Natural Area property boundary. The area was previously cropped to soybean (Glycine max), but the area has been taken out of cultivated agricultural production and had been managed as a prairie restoration for at least the last 15 years.
The Seidenstricker Prairie, which is presently an approximate 5-ha, fragmented tallgrass prairie remnant, is located in the Grand Prairie in Prairie County, Arkansas (Table 1; Figure 1). The prairie was once larger, but parcels of the prairie have been periodically converted to cultivated agriculture over the past 60 years.

Statistical Analyses
Based on similar procedures used recently by McKee et al. [5], a two-factor analysis of variance (ANOVA) was performed using the PROC GLIMMIX pro-

Results and Discussion
After 15 years of continued prairie function or agricultural land management, nearly all changes in soil chemical properties measured in the top 10 cm over time were affected by region, landuse, or both (Table 2). Changes in extractable soil Ca, Mg, and Zn over time varied between landuses within regions (P < 0.01), while changes in extractable soil S, Na, and Mn over time differed between regions (P < 0.01), and changes in extractable soil Fe and Mn over time differed between landuses (P < 0.02; Table 2).  and either landuse in the Ozark Highlands, which did not differ (Figure 2). Brye and Pirani [8] also reported that extractable soil Mg was generally greater under managed agriculture compared to native prairie in the Grand Prairie. In addition to adding Ca, agricultural lime materials often contain Mg as well if dolomite is used. Similar to Ca and Mg, extractable soil Zn increased over time (P < 0.05) under cultivated agriculture in the Grand Prairie, while extractable soil Zn did not change over time under managed agriculture or native prairie in the Ozark Highlands and under native prairie in the Grand Prairie (Figure 2). The change in extractable soil Zn was about 10 times greater under cultivated agriculture in the Grand Prairie (0.3 kg•ha −1 •yr −1 ) than under native prairie in the Grand Prairie and either landuse in the Ozark Highlands, which did not differ (Figure 2). Zinc is an essential plant micronutrient, thus Zn is periodically added while fertilizing row crops for optimal productivity.
Averaged across landuse, extractable soil S increased over time (P < 0.05) in the Ozark Highlands, but was unchanged over time in the Grand Prairie ( Table  3). The change in extractable soil S over time was nine times greater in the Ozark Highlands than in the Grand Prairie (Table 3). As a natural component of SOM, soil S likely increased as a result of the numeric, though non-significant, increase in SOM content in the Ozark Highlands, while SOM content significantly decreased in the Grand Prairie as reported by McKee et al. [5] over the same 15-yr time period and same treatments as in the current study. In contrast to S, averaged across landuse, extractable soil Na and Mn increased over time (P < 0.05) in the Grand Prairie, but were unchanged over time in the Ozark Highlands (Table 3). The change in extractable soil Na and Mn over time were 6.4 and 10 times greater, respectively, in the Grand Prairie than in the Ozark Highlands ( Table 3). Much of the Grand Prairie region of east-central Arkansas is characterized by soils in aquic soil moisture regimes (Table 1) and with relatively shallow water tables. Considering that Mn is prone to oxidation-reduction reactions as a result of water-logging and the development of anaerobic/reducing conditions, the periodic water-table fluctuations during wet seasons are likely responsible for the dissolution and re-precipitation of Mn near the soil surface in the Grand Prairie. In contrast, water tables in the Ozark Highlands tend to be deeper than in the Grand Prairie, which characterizes the udic soils sampled in the In contrast to all other measured extractable soil nutrients, changes in extrac- both nutrients tend to be applied to meet crop needs for a given year and not often applied in excess of crop needs to build up P or K in the soil for the future.
Alternatively, SOM decomposition would be the natural source of soil nutrients in the absence of fertilization, such as in a native prairie, or when fertilization rates only meet crop nutrient requirements for a given year, such as under agricultural management. However, over the same 15-yr time period and same treatments as in the current study, McKee et al. [5] reported that, averaged across native prairie and agricultural landuses, SOM contents in the top 10 cm decreased over time, which likely at least partially explains the decline in soil P, K, and Cu over time.
McKee et al. [5] also reported that soil bulk density in the top 10 cm did not  [5] reported that soil pH changed over time in

Conclusions
Near-surface extractable soil nutrients (i.e. P, K, Ca, Mg, S, Na, Fe, Mn, Zn, and Results from this long-term study demonstrate that 15 years is likely ample time to assess actual temporal changes in near-surface soil properties and to minimize the potential for mistakenly concluding that temporal changes in soil properties are attributed to reasons other than simple inter-annual variability.
Results from this field study also clearly demonstrate how differences in landuse and regional soil characteristics can affect near-surface soil nutrient contents, which should be taken into consideration when implementing conservation and/or ecosystem restoration activities.