Biodiversity genomics

A planting audit counts stems and species; a population survives on its genetic diversity. The distinction is not academic: stands that look successfully restored can be demographic sinks of near-clonal material that will fail at the first novel pathogen or drought. The procurement-level treatment of reference genomes and sequencing platforms is a separate subject; this page is about why genetic diversity is the quantity being restored.

Why heterozygosity is the currency#

Fitness depends on heterozygosity. In a small population, mating among relatives exposes recessive deleterious alleles — inbreeding depression shows up as poor germination, low fertility, reduced seedling survival, none of which a stem census records. Meanwhile genetic drift discards alleles at random, and the loss rate is set not by the number of individuals but by the effective population size: the count of breeders actually contributing genes, which in managed or restored stands is almost always far smaller than the census because founder stock is few mothers, a narrow seed collection, a single nursery lineage. Classical rules of thumb capture the asymmetry: a few dozen effective breeders keep short-term fitness roughly intact; hundreds are needed to preserve the raw material of adaptation over long horizons. And the loss is one-directional on human timescales — alleles erased by a bottleneck return only by mutation (effectively never) or by managed gene flow.

Why species counts mislead#

Two sites can hold the same species list while differing in whether each population is self-sustaining or a sink maintained by continuous immigration. A count-based audit rewards planting many species densely; a genetic audit asks whether the establishing populations carry enough allelic diversity — and the right diversity. Seed provenance is part of this: populations of a species are locally adapted, so seed imported from a warmer region may establish well and then fail at the climatic margin, while mixing too-distant provenances risks diluting local adaptation. Species-level survey methods cannot see any of this; the unit of interest sits below the species boundary.

What genomic monitoring actually measures#

Population-genomic data give quantities a census cannot: heterozygosity and allelic richness, inbreeding coefficients read from runs of homozygosity, effective population size estimated from the data rather than assumed, and parentage analysis showing whether planted trees actually reproduce or the stand is sustained by repeated replanting. Genetic rescue — introducing unrelated individuals into an inbred population — reliably restores fitness and remains underused because it looks like contamination of local purity. Environmental DNA extends detection to whole communities, but its scope is detection, not abundance: sequence counts in an eDNA sample are shaped by primer bias and by the completeness of the reference barcode database, so a species absent from the library is absent from the result regardless of its presence at the site. That calibration dependence is the same error budget that governs field-scale biomonitoring — see forest and agroecosystem biomonitoring.

The limiting quantity in restoration genetics is effective population size and the connectivity that maintains it. Diversity is lost within a few generations and rebuilt only by deliberate gene flow — which is why it belongs on the design page of a project, not in its monitoring appendix. How the degraded system got to where genetics matters at all is the subject of ecosystem restoration.