Phytosociology of a Bengaluru Woodland Plot: A Nested-Quadrat Survey with Diversity Indices

What this study does. A patch of urban woodland is not a random pile of plants: it has a structure — which species dominate, how they are distributed, and how many of them there are — and that structure can be measured with the classical tools of phytosociology. This project surveys two contrasting plots on the Bengaluru outskirts: a less-disturbed wooded plot (college-adjacent grove or protected fringe woodland) and a disturbed roadside scrub plot, using ten nested quadrats per plot (1 × 1 m for herbs, nested 5 × 5 m for shrubs, 10 × 10 m for trees) laid along transects with randomised starts.

From the species counts and cover the two-member team computes the full classical set: frequency, density, abundance, relative values and the Importance Value Index (IVI) for every species, plus the modern diversity summaries — Shannon–Wiener H′, Simpson's index and Pielou's evenness — and a species–area curve to show whether ten quadrats were enough. The plots are compared with a Jaccard similarity coefficient and a bootstrap confidence interval on the difference in Shannon index, so the ecological conclusion (disturbance simplifies the community, or does not) rests on numbers, not impressions.

Species that cannot be named in the field are pressed as herbarium vouchers and identified with regional floras, with doubtful identifications flagged honestly. The deliverables are the IVI tables, diversity indices, species–area curves and a written characterisation of both plots, produced from the team's own fieldwork.

Urban woodland remnants are being trimmed, grazed and encroached across Bengaluru's outskirts, and 'loss of green cover' is usually reported as area alone. Area is the crudest possible summary: what actually matters ecologically is which species the remnant still supports, which dominate its structure, and how its diversity compares with a less-disturbed reference — none of which follows from a satellite image.

Undergraduate field surveys rarely answer this properly: quadrats are thrown casually, sample size is never justified, species lists mix reliably and unreliably identified names, and the classical indices are computed once, without uncertainty and without connecting to an ecological question.

This project addresses the gap by surveying a protected and a disturbed plot with a defensible nested-quadrat design, computing the full phytosociological and diversity-index pipeline from the team's own counts, testing sampling adequacy with a species–area curve, and comparing the plots with a Jaccard coefficient and a bootstrap interval on the Shannon-index difference — so the ecological conclusion about disturbance is quantitative, reproducible and traceable to vouchers.

Abstract

This project characterises and compares the vegetation of a protected wooded plot and a disturbed roadside scrub plot near Bengaluru using nested-quadrat phytosociology. Ten quadrats per plot (herb 1 m², shrub 25 m², tree 100 m²) yield frequency, density, abundance and IVI per species; Shannon–Wiener H′, Simpson's index and Pielou's evenness summarise diversity; a species–area curve checks sampling adequacy; Jaccard similarity and a bootstrap interval on ΔH′ compare the plots. The study answers one question with field data: how does disturbance reshape community structure?

Introduction

Phytosociology — describing vegetation through the behaviour of species — remains the working language of Indian plant ecology, taught through the frequency–density–abundance trio culminating in the Importance Value Index. Urban fringes like Bengaluru's are natural laboratories: a single walk crosses wooded scrub, hedgerows and grazed, trampled road edges on similar soil and rainfall. Which species hold each plot together, and what diversity is lost across the boundary, is the question the classical indices exist to answer.

Existing knowledge and the gap

Textbooks define the indices meticulously but illustrate them with pre-cooked tables. Student surveys stop at a species list or a single ill-explained 'diversity' number, rarely justify quadrat size or number, almost never report sampling adequacy, and treat disturbed-versus-protected differences as obvious rather than measured. The Botanical Survey of India's flora resources make correct identification achievable for an undergraduate team, removing the excuse of unidentified 'grass sp.' entries swamping the data.

The gap this project fills: (a) a nested-quadrat design with a stated justification and a species–area curve showing whether ten quadrats were enough; (b) the complete IVI pipeline per layer, each plot's dominants named and defended; (c) diversity comparison with uncertainty — a bootstrap interval on the Shannon-index difference, not just two bare numbers; and (d) vouchers for every unfield-identified species.

Proposed study

1. Select and describe the two plots (area, aspect, soil surface, disturbance history); record coordinates and photographs.
2. Lay one transect per plot; place ten quadrats at fixed intervals with a randomised offset; record species counts and cover per layer.
3. Press vouchers of unfield-identified species; identify with regional floras and flag confidence.
4. Compute per-plot, per-layer tables: frequency %, density, abundance, relative values, IVI; IVI-ranked dominants.
5. Compute Shannon H′, Simpson's index and Pielou's evenness per plot; species–area curve per plot; Jaccard similarity; bootstrap 95 % CI on ΔH′ (2,000 resamples).
6. Interpret: which species dominate each plot, how disturbance changes structure, what the similarity coefficient says about shared flora.

Feasibility

• Technical: the survey needs quadrat frames, a herbarium press, field gear and a laptop with R (vegan) or Python — all free or already available in the department.
• Economic: transport and stationery only — under ₹500 apart from the press, which the department owns.
• Operational: each plot survey is one full day, repeated once; spare days absorb weather. Permission for the protected plot is arranged in Week 1; field days are scheduled post-monsoon, when herbs are identifiable but not drowned.

In scope

• Two plots, ten nested quadrats each, three vegetation layers (herb, shrub, tree), one survey season with a repeat visit.
• Classical phytosociological analysis (frequency, density, abundance, IVI), diversity indices (Shannon, Simpson, evenness), species–area curve, Jaccard similarity and bootstrap comparison of H′.
• Herbarium vouchers for unfield-identified species, identified with regional floras and confidence flags.

Out of scope

• Soil chemical analysis, soil-seed-bank study or faunal elements — a separate project each.
• Remote-sensing or NDVI cover estimation; this is a ground survey.
• Ordination methods (PCA, DCA) and environmental-gradient modelling — listed under future scope.
• Lichens, mosses and other cryptogams beyond a presence note.

Research design: a two-plot comparative field survey with systematic-random quadrat placement, analysed with classical phytosociological indices and modern diversity statistics. One factor (disturbance status) is contrasted descriptively and inferentially (bootstrap interval on ΔH′).

Plot selection (Weeks 1–2): two plots of roughly similar size (≥ 0.5 ha each) within 3 km of each other: Plot A — less-disturbed wooded grove or protected fringe (no lopping recent, closed canopy patches); Plot B — roadside scrub with visible grazing/trampling/lopping. Record GPS coordinates, aspect, slope, soil-surface notes, disturbance evidence and photographs; obtain written permission from the owner institution/authority.

Quadrat layout: one 50 m transect per plot placed along the plot's long axis; ten quadrat stations at 5 m intervals with a randomised starting offset (random number table); at each station the nested frames give the 1 m² herb quadrat, a 5 × 5 m shrub quadrat and a 10 × 10 m tree quadrat centred on the station. Herbs: count of individuals (or tillers/tufts for clonal species, noted) + visual cover class; shrubs: count of rooted stems + height class; trees: count of individuals ≥ 10 cm girth at breast height + girth measured with tape.

Identification: field identification with regional floras; anything uncertain is pressed (labelled with plot, quadrat, date) and identified in the department with flora keys; confidence flag (high/medium/low) recorded per species. Herbarium sheets follow standard pressing and mounting practice.

Analysis plan: frequency % = (quadrats containing species / total quadrats) × 100; density = total individuals / total quadrats studied (per layer); abundance = total individuals / number of quadrats of occurrence; relative values as percentages; IVI = relative frequency + relative density + relative abundance (tree layer, the classical use). Diversity: Shannon H′ = −Σpᵢ ln pᵢ, Simpson 1−D, Pielou J = H′/ln S, computed in R with vegan (or Python) using the pooled herb-layer counts per plot; species–area accumulation curve (mean of randomised quadrat orders, 100 permutations) to assess adequacy; Jaccard coefficient between plot species lists; ΔH′ bootstrap from quadrat-level resampling (2,000 resamples). All computations are run twice — once in R/Python, once in Excel for the key indices — as a cross-check.

Timeline (14 weeks, team of two): Weeks 1–2 site selection, permissions, flora familiarisation; Week 3 trial run of the quadrat protocol; Weeks 4–6 Plot A survey and voucher pressing; Weeks 6–8 Plot B survey and vouchers; Weeks 9–10 identification sessions; Weeks 11–12 analysis and figures; Weeks 13–14 report, slides and viva.

The study is a field-to-table pipeline: survey, vouchers, index computation, comparison. Each stage leaves an artefact that enters the report.

flowchart TD
  A["Select and describe Plot A (protected) and Plot B (disturbed)"] --> B["Written permission, GPS, photographs"]
  B --> C["Transect + 10 nested quadrats per plot (randomised offset)"]
  C --> D["Herb 1 m2: counts + cover"]
  C --> E["Shrub 5 x 5 m: stems + height class"]
  C --> F["Tree 10 x 10 m: girth at breast height"]
  D & E & F --> G["Species data sheets per quadrat"]
  G --> H["Unidentified species -> herbarium vouchers"]
  H --> I["Identification with regional floras, confidence flags"]
  G --> J["Frequency, density, abundance -> IVI per layer"]
  G --> K["R vegan: Shannon, Simpson, evenness, species-area curve"]
  J --> L["IVI dominants per plot"]
  K --> M["Jaccard similarity + bootstrap CI on delta H-prime"]
  L --> N["Floristic report and viva"]
  I --> N
  M --> N

Data organisation

One row per species per quadrat per layer, with plot, quadrat number and date — a tidy long-format sheet. All indices are computed from that single sheet, so a corrected count or a confirmed voucher name propagates through every table on re-run. Species names follow the flora's accepted name, with family noted and the voucher number quoted in the IVI tables.

Sampling adequacy check

flowchart LR
  A["Randomise quadrat order 100 times"] --> B["Mean cumulative species count per added quadrat"]
  B --> C{"Curve flattens by quadrat 8-10?"}
  C -->|Yes| D["Sampling adequate: report as is"]
  C -->|No| E["Add quadrats on a second visit; state the limitation"]

Results tables to fill (no numbers are pre-filled)

• Ordination: add soil pH, moisture and canopy-cover measurements per quadrat and run PCA/DCA to relate species pattern to environment.
• Seasonal comparison: repeat the survey in summer to quantify the herb layer's dry-season turnover.
• Regeneration study: seedling and sapling counts per tree species to test whether the protected plot's dominants are actually regenerating.
• Raunkiaer life-form and biological spectrum analysis across more plots of a disturbance gradient.
• Carbon proxy: allometric estimates of above-ground tree biomass from the girth data.
• A continuing departmental monitoring plot re-surveyed each batch, building a long-term floristic record.