The ATLAS calorimeter response to the electromagnetic component of a hadron shower is not equal to the response to the non-electromagnetic component. In general, the hadronic response is smaller than the electromagnetic response. This effect is mainly due to the energy lost in the breakup of nuclei or in nuclear excitation.
Jet Reconstruction Performance: From Atlas CSC note
For QCD cross-section measurements, prefer wider jets in order to capture the hard scattered parton kinematics completely. For W boson decaying into two jets or ttbar want narrow jets to distinguish jets in a busy environment.
Infrared safety: Problem for truth jets, 'Any soft particle not coming from the fragmentation of a hard scattered parton should not effect the number of jets produced.' No shared objects between jets leads to infrared safety.
Collinear safety: A jet should be reconstructed independent of the fact that a certain amount of transverse momentum is carried by one particle, or if a particle is split into two collinear particles. No seed in jet algorithm leads to collinear safety.
Topoclusters in Atlas are an attempt to reconstruct three-dimensional energy blobs. The shower overlap between the particles in the jet cannot be resolved as well in the central region (eta < 1.5) where the calorimeter cells are larger on the scale of the hadronic shower.
Local hadronic energy scale starts with a classification step characterizing clusters as electromagnetic, hadronic or noise based on location and shape. After this, cells inside hadronic clusters are weighted with functions depending on cluster location, energy and cell signal density. Corrections are derived from single particle simulations do not use the jet content.
H1 calibration is applied to towers or topoclusters (where local hadronic calibration has not been applied). The basic idea is that low signal densities in the cells indicates a hadronic signal and need correction and high density cells are more likely a electromagnetic shower and need no correction.
Jet mass reconstruction is difficult because calorimeter signals wash out the directions and energies of individual particles in the jet. Additionally, jet mass is best reconstruction if all of its original particles can be measured. This reconstruction is disturbed by undetected particles, like the ones curling in the solenoidal field or the ones losing too much energy in upstream inactive material.
Most jets at lower Pt in QCD di-jets are gluon jets. For higher Pt, a significant fraction of jets are produced from quarks. As gluons have a higher probability of radiating off other gluons than quarks have, one can expect more particles inside gluon jets and even more jets in the final state in the case of gluons.