Sites surveyed.

Two surveys of weed infestation were conducted in mid-April of 2015 and 2016 in 23 turfgrass areas in 10 districts or counties of Tianjin, a municipality in the North China Plain. The coordinates of the 23 surveyed sites and the functions of the turfgrass areas are presented in Supplemental Table 1. Festuca arundinacea Schreb. was the turfgrass plant at most sites; at two sites, it was the dominant turfgrass plant but occurred along with Poa pratensis L. The size of the area surveyed at each site was at least 1000 m². No fertilizer, mowing, or weed management programs had been carried out at the sites, and only some sites had been irrigated before the survey.

Data collection.

The plots were selected by walking across the central areas of the turfgrass sites and throwing a sample quadrat (0.25 m²) in a random direction. This resulted in a total of 15 sampling locations at each site. Our methods followed those of Andreasen et al. (1996). The weed species at each sampling location were identified and recorded for further analysis. To quantify and analyze the data, weed density, uniformity, frequency, relative density, relative uniformity, and relative frequency were calculated according to the following Eqs. [1]–[4] in reference to the methods of Thomas (1985, 1991):

View Expanded

where U_k is the field uniformity value for species k, X_ij is the presence (1) or absence (0) of species k in quadrat j (a quadrat is 0.25 m²) in field i, and n is the number of fields surveyed.

View Expanded

where D_k is the density of species k (plants/m²), D_ij is the number of plants in quadrat j in field i, and n is the number of fields surveyed.

View Expanded

where F_k is the frequency value for species k and Y_i is the presence (1) or absence (0) of species k in field i.

View Expanded

where RF_k, RU_k, and RD_k are the relative frequency, uniformity, and density of species k, respectively, and RA_k is the relative abundance of species k.

The severity of weed infestation in cropland depends not only on the abundance of the weeds but also on the weed coverage and height. To determine the dominance of different weed species in the weed community, two indexes, the dominance index (the product of the relative height of a weed to the turfgrass plant, scaling 0 to 1, and the relative coverage of that weed in a quadrat, scaling 0% to 100%, DI) and the percentage of dominance index (PDI) were used according to Eqs. [5] and [6] following the methods of Wu et al. (1997) and Zhou et al. (2004):

View Expanded

where DI_k is the dominance index of species k, RH_k is the relative height to the turfgrass plant of species k, and RC_k is the relative coverage in the quadrat of species k.

View Expanded

where PDI_k is the PDI value of species k. A weed species was defined as a dominant weed when PDI ≥10%, as an important weed when 1% ≤ PDI < 10%, and as a subdominant weed when PDI <1%.

Soil nutrient conditions (the contents of total nitrogen, hydrolyzable nitrogen, rapidly available phosphorus, rapidly available potassium, and organic matter) and soil physical properties (clay content and soil density) for the 0- to 25-cm soil layer were analyzed by the Tianjin Institute of Resource and Environmental Science. At sampling, the soil temperature of the 0- to 18-cm layer was measured with a digital probe thermometer (Kat. Nr. 30.1021; TFA Dostmann GmbH & Co. KG, Wertheim-Reicholzheim, Germany). The soil salinity of the 0- to 18-cm layer and the soil pH at the 0- to 11-cm layer were measured with PNT3000 and pH3000 m, respectively (STEP Systems GmbH, Nürnberg, Germany). The soil moisture of the 0- to 18-cm layer was measured with a Field Scout TDR 100 system (Item# 6440FS, Spectrum^® Technologies, Inc., Aurora, IL). The soil moisture data consisted of mass water content in 2015 and volumetric water content in 2016. Except for soil temperature and soil moisture (measured and analyzed separately in 2 years as two repetitions), the soil variables were measured in 2015 alone because they were not expected to change drastically in two successive years (Supplemental Table 1). Visual estimation of turfgrass deterioration was conducted during the survey. Deterioration ratings were based on a scale from 0% to 100% (where 0% = no turfgrass deterioration and complete coverage of the ground by turfgrass plants and 100% = severe turfgrass deterioration and either completely barren ground or groundcovered by other plants).

Statistical analysis.

Total weed density (TWD) and TDI (the sum of DI value that is the product of relative height and the relative coverage of each weed species) of all weed species in a 0.25-m² quadrat of each survey sites were calculated. To determine the correlations between the factors that might influence weed infestation and TDI in the turfgrass sites, PACA was performed using the software SPSS 19.0 (SPSS, Inc., Chicago, IL). Correlations between TDI and each of the 12 potential influencing factors (the abovementioned soil nutrient conditions, soil physical properties, temperature, moisture, salinity, pH, and turfgrass deterioration) for each survey site were calculated. PACA was used to determine the correlation between TDI and each of the 12 factors holding the effects of the remaining factors constant. Pearson correlation analysis was used to evaluate the correlation between two variables and two-tailed P values were calculated to indicate the significance at α = 0.05. Factors that were significantly correlated with TDI have influence on weed infestation. The data from 2015 and 2016 were analyzed separately because significant interactions of year with TDI, soil moisture, and other variables were found.

Weed community composition.

Across all turfgrass sites surveyed during early spring in both years, 37 weed species belonging to 14 different families were identified; these species are presented in alphabetic order in Supplemental Table 2. When classifying each weed according to life cycle, there were 17 perennial weeds, accounting for 45.9% of all of the weed species, and 20 annual or biennial weeds, accounting for 54.1% of the weed species. Asteraceae (16 species) and Brassicaceae (five species) were the dominant weed families, representing 43.2% and 13.5% of the weed species, respectively. When classified from the point of weed management, there were 35 broadleaved weed species and two grass weed species, accounting for 94.6% and 5.4% of the weed species, respectively.

Weed dominance analysis.

To quantify the dominance of each weed species surveyed across all sites over 2 years, the relative abundance value of each weed species (incorporating weed density, distribution uniformity, and frequency parameters) was calculated. The results from both years indicated that I. polycephala Cass., T. mongolicum Hand.-Mazz., I. japonica Thunb., H. lyrata Bge., T. pedunclaris (Trev.) Benth., C. hederacea Wall, L. apetalum Willd., P. asiatica L., Circus segetum Bge., and I. sonchifolia Hance. were the 10 most abundant weed species during the early spring, although their rank order slightly differed between the 2 years. All 10 weed species were broadleaved weeds and they belonged to five families: Asteraceae (six species), Boraginaceae, Brassicaceae, Plantaginaceae, and Convolvulaceae (one species each) (Tables 1 and 2).

Table 1.

The 10 most common weed species in areas of Festuca arundinacea Schreb. turfgrass during the early Spring of 2015.

View Table

Table 2.

The 10 most common weed species in areas of Festuca arundinacea Schreb. turfgrass during the early Spring of 2016.

View Table

In addition to relative abundance, the PDI of each weed (integrating weed coverage and height) was calculated to reflect the dominance of each weed in the community. The results indicated that I. polycephala Cass. and T. mongolicum Hand.-Mazz. were the two dominant weed species, with PDI values above 10% in both years.

Weed infestation variability and key influencing factors.

For each site, the TWD and TDI were calculated, and the dominant weed species were identified and summarized to identify the weed community and the differences in infestation severity among sites. The severity of weed infestation varied widely in both years. Both TWD and TDI were low at some survey sites (e.g., Chaoyang Avenue and Jingjin New City) and high at others (e.g., Jixian County and Xiaodian Park). The dominant weed species differed among the 23 survey sites and some were not among the 10 most common weed species at each site (Table 3).

Table 3.

Weed infestation and dominant weed species in 23 survey sites in the Tianjin region during the early Spring of 2015 and 2016.

View Table

Correlations between TDI and each of the 12 factors (including the soil nutrient conditions, soil properties, temperature, moisture, salinity, pH, and turfgrass deterioration) potentially related to weed infestation of turfgrass at each site were analyzed by PACA. The results indicated that the deterioration rate (D) of a turfgrass area was significantly and positively correlated with TDI (P = 0.001 in 2015 and P = 0.004 in 2016, with a partial correlation coefficient of 0.8238 in 2015 and 0.7648 in 2016) in both years. Soil organic matter and soil salinity (S) were negatively correlated with TDI, and the correlations were significant in 2015 (P < 0.05). However, no significant correlation between TDI and the remaining nine factors was observed in either year (Table 4).

Table 4.

Partial correlation analysis results of weed infestation and 12 related factors in the Tianjin region during the early Spring of 2015 and 2016.

View Table

Weed community composition.

When weeds appear in a crop production system, identification of the weed community composition is required before economical and effective management practices can be undertaken (Frick and Thomas, 1992; McCully et al., 1991; Thomas, 1985, 1991). The results of our survey indicated that perennial and biennial broadleaved weed species belonging to the families Asteraceae and Brassicaceae are the weeds to target during early spring in the Tianjin region, China. Severely harmful annual weed species, e.g., large crabgrass [Digitaria sanguinalis (L.) Scop.] and small crabgrass (Digitaria ischaemum Schreb. ex Muhl.) had not infested by mid-April in this region. The absence of these species might be due to unsuitable environmental conditions for weed infestation (e.g., the combination of soil temperature and moisture) during this period (Masin et al., 2005).

Festuca arundinacea Schreb., as a popular cool-season turfgrass plant, is planted extensively in large areas of China. The present findings of weed community composition in turfgrass areas during the early spring in Tianjin, a typical region of the North China Plain, are of broad interest to turfgrass researchers and managers from other regions where F. arundinacea Schreb. is planted.

Common and dominant weed species.

Weed density, frequency, uniformity, and their relative values to the whole community are commonly used parameters to analyze weed composition in crop production systems. However, each parameter has limitations when used independently because it might not accurately reflect the ecological importance of a species within a community (Andreasen and Stryhn, 2008; Nkoa et al., 2015; Thomas, 1985; Uddin et al., 2010). Hence, we integrated weed density, field frequency, and field uniformity into relative abundance, and we integrated relative height and coverage into PDI for each weed species. The top 10 weed species ranked according to PDI value were the same as those ranked according to relative abundance value (Tables 1 and 2). This demonstrated that these two parameters are appropriate for determining the severely infesting weed species, and our ranking of the 10 most common weed species we identified during this period in this region is highly credible.

Based on the analyses of PDI for the 2 years, I. polycephala Cass. and T. mongolicum Hand-Mazz. were identified as the two dominant weed species. Esmaili and Salehi (2009) and Ferguson et al. (2016) reported that Taraxacum officinale is problematic for turfgrass managers each year, and this plant strongly competes with turfgrass plants for resources (Franssen and Kells, 2007). Although T. mongolicum Hand-Mazz. is a different species from the congeneric T. officinale and although there are few reports on I. polycephala Cass infestation of turfgrass areas, infestation by the two weed species identified here warrants attention.

Key factors influencing weed infestation.

Weed populations in turfgrass areas are governed by both ecological and human activities, and the interactions that invariably occur between weed species and environmental or management variables are very complex (Nkoa et al., 2015). According to the PACA in our study, the deterioration rate of turfgrass was significantly and positively correlated with TDI in both years. Most of the turfgrass areas with high deterioration rates were severely infested by weeds and located in residential areas (e.g., Xishuangtang Park and Xiaodian Park). This observation is consistent with the results of Uddin et al. (2010), who demonstrated that residential lawns had more weeds than did golf course putting greens, athletic fields, and sod farms. High daily foot traffic, low daily maintenance, and aged turfgrass plants are likely reasons for this phenomenon. Soil organic matter typically influences weed abundance in the field (Dieleman et al., 2000a, 2000b; Medlin et al., 2001). The negative correlation observed between SOM and TDI in our study might be due to the positive effects of high SOM on turfgrass growth. It precludes severe weed invasion and establishment. Similar to SOM, soil salinity showed a negative correlation with TDI (e.g., the soil salinity at the Chaoyang Avenue site was high, but the weed infestation severity was low). This result might be due to inhibitory effects of high soil salinity on the growth of most of the weed species and the subsequent reduction of the weed infestation of turfgrass. No significant correlation between TDI and any of the remaining nine parameters was observed; it is possible that the differences in these parameters among the different turfgrass areas were too small to be detected statistically.

Our 2-year survey identified the 10 most common and the two most harmful weed species in F. arundinacea Schreb. turfgrass areas and the three key factors (turfgrass deterioration rate, SOM and salinity) influencing weed infestation of these areas during early spring in the Tianjin region, China. These findings allow turfgrass researchers and managers to target specific harmful weed species. In addition, they can improve the development of preventive and curative weed management programs and promote precise and efficient weed control. By understanding the key factors that influence weed infestation, nonchemical strategies to control harmful weeds can be adopted, such as optimizing daily turfgrass maintenance (e.g., mowing, irrigation, fertilization, and other practices), restoring areas of turfgrass deterioration and controlling weeds by using saline water (see Uddin et al., 2014).