Given the extent, some of the heteroplasmic SNPs should be genuine (see comment above on the used polymerase). When looking at the supplement tables reporting the haplotypes exemplifying the heteroplasmy, one thing strikes the eye: they are all satellite haplotypes of the reference plastome sequence.

If common biparental inheritance would be the cause of the heteroplasmy, one would expect some more haplotype structuring as one can observe within wide-spread species or species aggregates and analogous to what we can find e.g. in multicopy nuclear gene regions. In this context, it should be interesting to map the detected heteroplasmic haplotypes across a larger range.

If lineage crossing contributed to heteroplasmy, one should find evolutionary distinct haplotypes within individuals, not only satellite haplotypes.

At this point, a simple alternative explanation for the findings would be that since there is more than one plastid ring within a chloroplast and the oocyte carries one or more chloroplasts, what one observes here by closing in with highest resolution methods is just within-individual variation due to stochastic replication errors some of which are possibly passed on to the next generation.

So far, the documented heteroplasmy doesn't ".. suggest[s] that intra-individual heterozygosity in the G. tongolensis plastome may be observed in functionally paralogous copies within an individual caused by biparental plastid inheritance." – functionally paralogous copies should be substantially different (sequence-wise) going far beyond satellite haplotypes. Do you have any evidence that at least the common heteroplasmic SNPs do change the amino-acid chain in a way, the proteins could not perform the same function?

PS Just because genes (or genomes) are inherited from both parents, they do not necessarily differ in function. In fact, many cases of documented intra-genomic variation due to biparental inheritance or hybridisation relate to functionally homologous (orthologous or homeologous) copies.