We present a study of comet C/2017 K2 ( PANSTARRS ) using prediscovery archival data taken from 2013 to 2017 .
Our measurements show that the comet has been marginally increasing in activity since at least 2013 May ( heliocentric distance of r _ { \mathrm { H } } = 23.7 AU pre-perihelion ) .
We estimate the mass-loss rate during the period 2013–2017 as \overline { \dot { M } } \approx \left ( 2.4 \pm 1.1 \right ) \times 10 ^ { 2 } kg s ^ { -1 } , which requires a minimum active surface area of \sim 10–10 ^ { 2 } km ^ { 2 } for sublimation of supervolatiles such as CO and CO _ { 2 } , by assuming a nominal cometary albedo p _ { V } = 0.04 \pm 0.02 .
The corresponding lower limit to the nucleus radius is a few kilometers .
Our Monte Carlo dust simulations show that dust grains in the coma are \gtrsim 0.5 mm in radius , with ejection speeds from \sim 1–3 m s ^ { -1 } , and have been emitted in a protracted manner since 2013 , confirming estimates by Jewitt et al .
( 2017 ) .
The current heliocentric orbit is hyperbolic .
Our N-body backward dynamical integration of the orbit suggests that the comet is most likely ( with a probability of \sim 98 % ) from the Oort spike .
The calculated median reciprocal of the semimajor axis 1 Myr ago was a _ { \mathrm { med } } ^ { -1 } = \left ( 3.61 \pm 1.71 \right ) \times 10 ^ { -5 } AU ^ { -1 } ( in a reference system of the solar-system barycentre ) .