TABLE 3.

Estimation of error for the parametric models used to describe the OTU abundance distribution in each community and estimates of OTU richnessa

ModelMicrobial groupSoil typeModel errorEstimated OTU richness
Power lawBacteriaDesert15.26 × 103
Prairie1.502 × 104
Rainforest4.502 × 104
ArchaeaDesert66.4>1 × 107
Prairie17.32 × 103
Rainforest21.31 × 103
FungiDesert30.02 × 104
Prairie20.52 × 109
Rainforest9.662 × 103
VirusesDesert1.081 × 103
Prairie0.622 × 104
Rainforest7.6>1 × 108
Log-normalBacteriaDesert14.86 × 104
Prairie1.904 × 105
Rainforest5.32>1 × 106
ArchaeaDesert1043 × 105
Prairie20.63 × 103
Rainforest25.23 × 103
FungiDesert42.77 × 104
Prairie23.7>1 × 106
Rainforest9.002 × 104
VirusesDesert1.081 × 103
Prairie0.661 × 105
Rainforest10.4>6 × 106
LogarithmicBacteriaDesert11.32 × 103
Prairie5.704 × 103
Rainforest13.44 × 103
ArchaeaDesert1421 × 103
Prairie23.22 × 102
Rainforest36.32 × 102
FungiDesert62.02 × 103
Prairie29.02 × 103
Rainforest28.11 × 103
VirusesDesert1.081 × 103
Prairie0.915 × 103
Rainforest3.0>1 × 106
  • a The best descriptive function of the community structure is the one that minimizes the calculated error (shown in boldface). The results for the exponential model are not shown because this model had the highest model error in all 12 cases.