Differences in distribution and community structure of plant-parasitic nematodes in pecan orchards between two ecoregions of Georgia

Working in conjunction with Cooperative Extension agents and pecan growers, PPNs were sampled in 6 and 22 commercial pecan orchards located in four and 21 Georgia counties from Piedmont and Coastal Plain regions, respectively (Tables 1–3). Actively managed pecan orchards were selected from each county with possible symptoms (stunted growth and loss of vigor) of PPN infestations. In each orchard, 7–14 individual pecan trees were arbitrarily selected for soil sampling.

Each tree was then sampled by removing 10 random soil cores (15 cm deep x 2.5 cm diam) under canopy drip line around tree stem (Khanal et al., 2016). Soil cores from each tree were combined into one composite sample and 7–14 such composite samples were collected for each orchard. A total of 282 composite samples collected from all the selected pecan orchards and each composite sample was placed in a plastic bag and transported back to the Extension Nematology Laboratory, Athens, Georgia in coolers to assay nematode populations.

Plant-parasitic nematodes were collected from a 100 cm³ soil sub-sample taken from each of 282 composite samples using centrifugal sugar floatation technique (Jenkins, 1964). Nematodes from each sample were identified to genus level using diagnostic keys by Mai et al. (1996) and counted at 40X magnification using an inverted compound microscope. Then their frequency of occurrence, abundance, and maximum population densities were calculated using standard formulas (see Tables 2 and 3 captions).

Soil pH and texture (proportions of sand, silt, and clay) analyses were conducted with four samples from two representative pecan orchards (Clarke and Walton counties – Piedmont ecoregion) and with six samples from four pecan orchards (Dougherty, Houston, Peach, and Tift counties – Coastal Plain ecoregion). A total of 10 soil samples were analyzed at the University of Georgia soil test laboratory in Athens, Georgia using standard methods (Gee and Bauder, 1986; Thomas, 1996).

We used a trio of multivariate analyses to evaluate the potential for ecoregion associated pecan PPN community differences. First, we used Non-Metric Multi-Dimensional Scaling (NMS or NMDS) (Kruskal, 1964; McCune and Grace, 2002), an ordination analysis that more accurately represents the structure within and between biological communities that PCA and PCoA (McCune et al., 2002), to visualize the patterns of PPN abundance and occurrence in soil samples collected from the Piedmont and Coastal Plain. NMS analysis was conducted using two data matrices, a primary and a secondary matrix. The primary matrix consisted of count data (relative abundance) from all 11 nematode from each pecan tree composite sample. Because nematode relative abundance varied by several orders of magnitude between genera and some genera were not present in a composite sample (a zero), we ln (x + 1) transformed the count data to conserve relative abundance (adding a 1 to all values enables us to ln transform all data while maintaining the zero in the data; ln 1 = 0, ln 0 is undefined and not appropriate for statistical analyses) and conform to multi-variate normality. Data from the primary, ln x + 1 transformed matrix, were converted to Gower’s distance values (Gower, 1971) for ordination. These values index the relative distance between the nematode counts for each pecan tree by simultaneously accounting for both PPN presence and relative abundance. NMS produces a scatterplot of how the pecan trees are related to each other given the Gower distance of each composite sample and calculates the centroid for each PPN genus in multivariate space. A secondary matrix was used to group the pecan tree PPN samples by ecoregion, either Piedmont or Coastal Plain. In ordination analyses, the more distant or closer the points are to each other in ordination space, the more dissimilar or similar the PPN communities in those samples are relative to each other, respectively. If there are differences in the PPN communities associated with ecoregion, we expect the NMS ordination to group pecan composite samples by ecoregion. NMS was accomplished through 500 real runs and 500 randomized runs in the program PC-ORD 7 (McCune and Mefford, 2016).

To test for statistically significant differences between the Piedmont and Coastal Plant pecan PPN communities, we retained the primary matrix (ln x + 1 transformed) and secondary matrix in the NMDS analysis and used them both for Multi-Rank Permutation Procedure (MRPP) (Mielke and Berry, 2001). MRPP is similar to an ANOVA but it is used for multivariate data. It evaluates group membership by comparing the distance values from each sample in the a priori selected groups (Piedmont and Coastal Plains ecoregions), calculates a test statistic for the between group differences (T, analogous to a student’s t), and indexes the effect size though a chance agreement within groups (A-statistic). A p-value for between groups statistical significance is obtained through the T-statistic (Mielke and Berry, 2001). MRPP was run from Gower’s Index values (Gower, 1971) in the program PC-ORD 7.

Because MRPP only tests for group membership, e.g. whether the pecan PPN communities differed between Piedmont and Coastal Plain ecoregions, we applied Indicator Species Analysis (ISA) (Dufrêne and Legendre, 1997) to identify which PPN genera, if any, were statistically associated with either ecoregion. Indicator values range from 0 to 100, with a 100 indicating a perfect indicator species (one that is mutually exclusive to a group and always occurs in the highest relative abundance within that group) and zero (a species that is not affiliated with any group either through occurrence or relative abundance) (see Severns and Sykes, 2020 for a description of the analysis). Probability values of ecoregion association were attained through 5,000 randomizations in PC-ORD 7.

The distribution of 11 PPN genera in six and 22 pecan orchards varied among the four and 21 counties that located in Piedmont and Coastal Plain ecoregions, respectively (Table 1). Also, the percent frequencies of occurrence, abundances, and maximum population density of five major PPNs including Meloidogyne, Mesocriconema, Paratrichodoru, Helicotylenchus, and Xiphenema that occurred on pecans differed between the nematodes and counties that are located in two different ecoregions of Georgia (Table 2). For example, only Mesocriconema was found in all locations in both ecoregions with its percent frequency of occurrences ranging from lowest (30%) to highest (100%). Of the four counties surveyed in the Piedmont region, PPN genera including Meloidogyne, Xiphenema, and Helicotylenchus found in all four counties with lowest (10%) and highest ( > 70%) of frequencies of occurrence but Paratrichodorus found in only three counties with only 10–23% of frequencies of occurrence (Table 2). Of the 21 counties surveyed in the Coastal Plain region, PPN genera including Paratrichodorus, Meloidogyne, Xiphenema, and Helicotylenchus found in 20, 19, 16, and 15 counties, respectively with lowest (10%) and highest (> 70%) of frequencies of occurrence (Table 2).

The distribution of 11 PPN genera that were associated with pecans in both geographic regions of Georgia varied between Coastal Plain and Piedmont ecoregions (Table 3). Also, the differences in frequencies of occurrences of PPNs and their communities are correlated with the properties of soil from both ecoregions (Table 3). For example, the percent frequencies of occurrence of ring (93%) and root-knot (44%) nematodes were comparatively higher in Coastal Plain, which is dominated with sandy soil containing average of 87.52% sand, 5.92% silt, and 6.56% clay with a pH of 6.6 (averages of six samples) than those in the Piedmont ecoregion (75 and 37%, respectively), which is dominated with sandy-clay-loam soil containing average 65.47% sand, 17.25% silt, and 17.28% clay and a mean pH of 5.5 (averages of four samples). In contrast, the percent frequencies of occurrence of dagger (33%) nematodes was higher in the Piedmont than in the Coastal Plain ecoregion (28%) (Table 3). The maximum population densities of nine nematodes were comparatively higher in Coastal Plain with sandy soil than those in the Piedmont ecoregion with sandy-clay-loam soil (Table 3).

NMDS ordination of the Georgia pecan PPN community yielded a three-dimensional solution that explained 91.9% of the variation in the soil sample data. The final stress of the 3-d solution was 13.36, a relatively stable NMDS ordination for ecological communities (McCune and Grace, 2002). Composite soil samples from the Piedmont ecoregion occurred primarily on one side of the ordination (left), while samples from the Coastal Plain occurred primarily on the opposite side (Fig. 1). In the Piedmont, the centroids (asterisks in Fig. 1) of lesion and sheath nematodes occurred in the center of the Piedmont samples (dark gray pyramids), suggesting these PPNs were proportionately dominant in the Piedmont. The centroids for stubby root, dagger, and stunt were positioned within the Coastal Plain samples (dark gray pyramids), while the centroids for ring, RKN (root-knot nematode), cyst, and spiral were not as obviously associated with either ecoregion (Fig. 1).

Figure 1:

MRPP indicated that the pecan PPN communities statistically differed between the two ecoregions (MRPP results: T = 3.083, A = 0.043, observed delta value = 0.115, p < 0.000000001). Indicator Species Analysis (ISA) identified that ring and stubby-root nematodes were statistically associated with Coastal Plain samples and that spiral, sheath, and lesion nematodes were statistically associated with the Piedmont ecoregion samples (Table 4 ISA). The PPNs that the Indicator Species Analysis identified as being statistically associated with either the Piedmont or Coastal Plain ecoregions are likely those that contributed to the statistical differences of PPN communities between ecoregion indicated by MRPP.