The past several decades have witnessed extraordinary developments in the area of naturally sourced functional nanomaterials, including nanocelluloses. Cellulose nanofibrils (CNFs) and cellulose nanocrystals (CNCs) have been obtained by downsizing higher-ordered cellulose fibers from different sources like hard- and softwood and agricultural biomass. A critical success factor of the industrialization of nanocelluloses has included optimizing the isolation techniques at both the commercial and designer levels. The coupling of mild chemical pretreatment with post-treatment mechanical processing has been the most promising way to obtain nanocelluloses economically. Specifically, the 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO)-mediated oxidation method has been the primary pretreatment technique used in academia and industry. Nevertheless, TEMPO remains expensive, corrosive, and toxic. Thus, crucial and emerging challenges of CNF production include stoichiometric optimization of chemicals and better alternatives to the status quo. Today, there are numerous approaches to prepare nanocelluloses, each affording specific chemistry and dimensions. In this thesis, two additional carboxylation routes were applied, including sequential periodate-chlorite oxidation and carboxymethylation, which have been available in the literature but have remained subsidiary to TEMPO and without good reason. All involved heterogeneous reactions on cellulose introducing negatively charged surface carboxylates. The charges aided in fiber defibrillation into CNFs and maintained their homogeneity as electrostatically stabilized aqueous dispersions. Non-wood cellulose feedstock has also been understudied. Thus, rice straw and almond hulls, as agriculture biomasses produced in significant amounts in California and worldwide, served as the supply for cellulose fibers. The use of two cellulose sources and three surface modification processes afforded analysis of both source-linked and process-linked quality attributes of CNFs. For example, oxidation of rice straw cellulose with regioselective TEMPO/NaClO at the primary C6 position gave a quasi-isotropic CNF cross-section. Conversely, C2, C3 regioselective periodate-chlorite oxidation and non-regioselective carboxymethylation routes produced CNFs with anisotropic lateral dimensions or dominant hydrophilic widths to hydrophobic heights and more significant electrostatic and steric effects on the interfaces of CNFs. One interest was in the tendency of CNFs to physically entangle into gels at low <1 w/v% aqueous dispersion, which was advantageous for the macroscopic assembling of CNFs into aerogels. Three-dimensional, high-porosity (>99.5%), and low-density (<8.0 mg/cm3) ice-templated aerogels were the medium for characterizing source- and process-related self-organizing and self-assembling behaviors and functionalities of CNFs. Results confirmed that regardless of source and process, CNFs maintained a high axial ratio and demonstrated Janus-type hydrophilic and hydrophobic surfaces proven by the amphiphilic super absorption of both polar and nonpolar liquids of the aerogels. The ice-template/freeze-dry formation of PC-CNFs and CMCNFs gave stiffer aerogels, displaying the highest toughness and the greatest strengths. This confirmed predominant CNF self-assembling along larger hydrophilic surfaces, demonstrating the importance of the lateral dimensions of CNFs on the mechanical performances of the assembled aerogels. Concerning source, applying the same TEMPO technique on almond hull produced nanofibrils with ultrahigh lateral and axial ratios and residual lignin content. The resulting aerogels exhibited enhanced mechanical performances linked to significant interfibril physical entanglements and lignin as a natural binding agent. As such, even a biomass feedstock of low cellulose content is a promising natural source for less pristine nanofibrils with value-added properties.