Large rivers are fundamental to human societies and consequently their ecosystems have come under increasing pressure from a range of developments and uses. Despite this, there is still a major knowledge gap understanding food webs supporting fisheries of large river ecosystems. Quantifying the contributions of carbon sources that support food webs is an important and growing field of ecological research, with implications for future management and rehabilitation of large rivers. I reviewed theoretical concepts addressing carbon flow in large river food webs where organic matter from floodplains (Flood Pulse Concept), local aquatic sources (Riverine Productivity Model), or leakage from upstream processing of terrestrial organic matter (River Continuum Concept) can fuel secondary production. Recent empirical evidence highlights the importance of autochthonous carbon, especially in the form of benthic algae and phytoplankton, to food webs in a variety of large rivers along with a range of secondary carbon sources that can assume importance depending on temporal and spatial variation in hydrogeomorphic conditions. The geographic spread of studies addressing carbon flow in large river food webs is steadily increasing, although information remains sparse on temperate Southern Hemisphere rivers and long-term data sets on carbon flow are generally lacking. I measured natural abundances of stable carbon (δ13C) and nitrogen (δ15N) isotopes to quantify spatial and temporal patterns of carbon flow through aquatic food webs in the lowland section of New Zealand’s longest river, the Waikato River. Zones of potential ecological importance influencing carbon transfer along the lower Waikato River were identified using a combination of (i) high-frequency, along-river water quality measurements collected during four seasons and (ii) river channel morphology data derived from aerial photos. A multivariate statistical approach was developed to identify three hydrogeomorphic zones shaped by the physical complexity and channel character of constituent river reaches, and characterised by shifts, sometimes transitional, of physico-chemical variables. Changes in water clarity, chlorophyll fluorescence and specific conductance were driven by tributary inflows and tidal influence. Carbon flow estimated using the mixing model IsoSource supported predictions of the Riverine Productivity Model, with autochthonous algae and biofilms (phytomicrobenthos) the most important basal carbon sources contributing to consumer biomass in all three zones. These sources were often supported by C3 aquatic macrophytes and allochthonous C3 riparian plants. However, the relative importance of organic carbon sources appeared to change depending on season and zone, likely in response to variations in water temperature and flow, particularly in the unconstrained zone of the lower river. It was also demonstrated that to draw robust conclusions, consideration must be given to quantifying the isotopic signatures of organisms lower in the food web, as these can change significantly between sampling times and hydrogeomorphic zones. Tributary confluences can be hotspots for biological production and provide novel carbon sources from donor sub-catchments in large river systems. Littoral food webs and water quality were compared between two main stem habitats (constrained and unconstrained hydrogeomorphic zones) and tributary junctions representing those fed by streams, lakes and wetlands during seasonal low flows when these habitats were likely to be most different. δ13C and δ15N isotopes were then employed using the Bayesian statistics R package Stable Isotope Analysis in R (SIAR) to estimate carbon flow through food webs and also to estimate measures of trophic structure. Pathways were also tested using analysis of fish stomach contents. SIAR mixing models confirmed that autochthonous benthic carbon was the most important carbon source to littoral food webs in all habitats. Riparian carbon appeared to be the most important secondary carbon source to fish consumers, and estimates of its contribution were often greater in tributary junctions compared to fish of the same species in the main stem. Trophic patterns of fish species collected in both the main stem and tributary junctions were similar amongst habitats, as were community metrics estimated using stable isotope signatures and SIAR. This study demonstrates that, while they may add to the lateral complexity of the riverscape, permanently connected habitats such as tributary junctions do not necessarily contribute to overall food web complexity. In this study tributary junctions tended to be steep-sided, and complex littoral habitats containing woody debris and macrophytes were typically rare, potentially limiting the development of more complex food webs. These results contribute to the ever-improving data regarding food web ecology in large rivers, particularly with regard to carbon flow, and the role played by lateral habitats and hydrogeomorphic zones in shaping these processes. This study also provides information and recommendations that provide direction for future research and management actions aimed at aiding the rehabilitation of the lower Waikato River, its riverscape and biological communities.