1: \begin{abstract}

2: Perturbative QCD (pQCD) running coupling $a(Q^2)$ ($\equiv \alpha_s(Q^2)/\pi$)

3: is expected to get modified at low spacelike momenta 

4: $0 < Q^2 \alt 1 \ {\rm GeV}^2$ so that, 

5: instead of having unphysical (Landau) singularities 

6: it remains smooth and finite there, due to infared (IR) fixed point.

7: This behavior is suggested by: Gribov-Zwanziger approach,

8: Dyson-Schwinger equations (DSE) and other functional methods, 

9: lattice calculations, light-front holographic mapping

10: AdS/CFT modified by a dilaton background, and by

11: most of the analytic (holomorphic) QCD models. 

12: All such couplings, $\A(Q^2)$, differ from the pQCD

13: couplings $a(Q^2)$ at $|Q| \agt 1$ GeV 

14: by nonperturbative (NP) terms, typically by

15: some power-suppressed terms $\sim 1/Q^{2 N}$.

16: 

17: Evaluations of low-energy physical QCD quantities in terms of such $\A(Q^2)$

18: couplings (with IR fixed point) at a level beyond one-loop are usually performed

19: with (truncated) power series in $\A(Q^2)$. We argue that such an evaluation is not correct, because the NP terms in general get out of

20: control as the number of terms in the power series increases.

21: The series consequently become increasingly unstable under the variation

22: of the renormalization scale, and have a fast asymptotic

23: divergent behavior compounded by the renormalon problem. We argue that an alternative series

24: in terms of logarithmic derivatives of $\A(Q^2)$ should be used.

25: Further, a Pad\'e-related resummation based on this series gives results which

26: are renormalization scale independent and show very good convergence.

27: Timelike low-energy observables can be evaluated

28: analogously, using the integral transformation

29: which relates the timelike observable with the corresponding

30: spacelike observable.

31: \end{abstract}

32: