Measurements collected during the AUSWEX field campaign, at Lake George (Australia), resulted in new insights into the processes of wind wave interaction and whitecapping dissipation, and consequently new parametrizations of the input and dissipation source terms. The new nonlinear wind input term developed accounts for dependence of the growth on wave steepness, airflow separation, and for negative growth rate under adverse winds. The new dissipation terms feature the inherent breaking term, a cumulative dissipation term and a term due to production of turbulence by waves, which is particularly relevant for decaying seas and for swell. The latter is consistent with the observed decay rate of ocean swell. This paper describes these source terms implemented in WAVEWATCH III and evaluates the performance against existing source terms in academic duration-limited tests, against buoy measurements for windsea-dominated conditions, under conditions of extreme wind forcing (Hurricane Katrina), and against altimeter data in global hindcasts. Results show agreement by means of growth curves as well as integral and spectral parameters in the simulations and hindcast.Keywords: wave modelling, wind input, negative input, whitecapping dissipation, swell dissipation 105 ("negative input") is that it is calculated from friction velocity and thus wind speed and does not dissipate swells in absence of wind (Tolman and Chalikov, 1996; Tolman, 2002). The input in TEST451 accounts for swell dissipation due to interaction with the air and thus can become negative (Ardhuin et al., 2010(Ardhuin et al., , 2011a. The dissipation includes a threshold and cumulative term. 110 The outline of the paper is as follows. Section 2 provides detailed description of the wind input term and the dissipation source terms implemented in WAVEWATCH. Section 3 contains description of the setup and results for the idealised academic tests and simulations selected to test the performance of the observation-based source terms. In Section 4 the results are discussed 115 and the conclusions are formulated in Section 5. 6 2. Source terms 2.1. Wind input The wind input function represents the energy flux transferred from wind to waves. This term is due to form drag, i.e. pressure acting on the surface 120 slope of the waves (e.g. Donelan et al., 2006). AUSWEX data analysis and the wind input parameterization reported by Donelan et al. (2005, 2006) and Babanin et al. (2007a) shows dependencies that have not been reported in previous experiments. Measurements of wave growth during AUSWEX were available for a range of wind-forcing conditions including very young waves 125 U 10 /c p = 5.1 − 7.6 (c p is the phase speed at the spectral peak) of varying steepness. This unique dataset revealed a number of new features: (i) full airflow separation with a relative reduction of wind input for conditions of strong winds/steep waves, if compared with its extrapolation from the moderate conditions (Donelan et al., 2006); (ii) dependence of the wave growth rate on 130 w...