Recent measurements of $e^{+}/e^{-}$ ratio in the 1-500 GeV range
were accurate enough to challenge the ``standard'' model of CR production
in SNRs. After a brief decline between $2-8$
GeV, the ratio rises steadily with a trend towards saturation around
200-400 GeV. These observations appear to be in conflict with the diffusive
shock acceleration (DSA) mechanism, operating in a $single$ SNR
shock, so most explanations appeal to multi-source scenarios. These,
however, have a significant number of adjustable parameters.

We argue that $e^{+}/e^{-}$ ratio is explicable by a single SNR-DSA
if $e^{+}$ are produced by shocks which: (i) propagate in clumpy
gas media, and (ii) are modified by accelerated CR \emph{protons}.
The protons penetrate into the dense gas clumps upstream to produce
positrons and, $charge~the~clumps~positively$. The induced electric
field expels positrons from the clump. These positrons are then accelerated
by the modified shock, thus developing a harder spectrum than that
of the CR electrons accelerated elsewhere. Mixing these $e^{+}$ and
$e^{-}$ populations explains the increase in the $e^{+}/e^{-}$ ratio
at $E>8$ GeV. Its decrease at $E<8$ GeV is due to a subshock weakening
that results from the shock modification. The subshock weakness also
explains why the neutral gas clumps survive its ionizing radiation.
Contrary to the positively charged secondary $e^{+}$ expelled from
the dense gas clumps and accelerated by the DSA mechanism, negatively
charged $e^{-},$ in part also $\bar{p}$ along with heavier positive
secondaries, such as boron, are locked in the clumps. The weakened
subshock engulfs the clumps, so the locked particles evade acceleration.
Scenarios for the 100-300 GeV AMS-02 fraction possibly exceeding the
model prediction, including possible dark matter and pulsar contributions,
are briefly discussed.