Future studies which investigate real-time measurements of GHG emissions over geospatial scales, should not just focus on agricultural landscapes but other ecosystems too. As research aimed at designing technologies to make accurate measurements of GHG emissions is not only relevant to agriculture but other ecological landscapes should too. For example, temperate grasslands in Europe are currently sequestering carbon (RIA, 2016). Wireless sensors would greatly aid assessing the rate of uptake and exchange of carbon in these soils over the next coming decades. This would be of huge benefit to policy planners (such as the Conference of Parties) who require carbon budgets to draw up plans on lowering GHG emissions. Likewise, temperate deserts such as those in Central Asia, are huge sinks of carbon, containing 1.34–34.16 Pg in the first 1m of top soil and an additional 10.42–11.43 Pg stored at deeper depths (1–3 m) (Li et al. 2015).
Had enough time been feasible to make more static gas chambers a WSN would have been established. WSNs play several roles in scientific and industrial fields, for example, measuring GHG emissions such as CO2 (Stocker et al. 2013; Metz et al. 2007), CH4 emissions (Tümer and Gündüz, 2010) and NH3 and N2O emissions from fertilizer (Jung et al. 2008; Ruiz-Garcia, et al. 2009). While wired sensors can perform similar tasks, wireless sensors can operate in a diversity of environments with advantages in cost, size, power, flexibility and distribution (Ruiz-Garcia et al. 2009). Furthermore, they can also access spaces that wired sensors cannot, for example concrete structures (Norris et al. 2008). In expanding this project, additional studies should try to develop a WSN involving Arduino, many closed system chambers and T6713 sensors.
Similarly, on a larger scale, a more complex system than Excel would be necessary to store and process greater amounts of sensor and GC data. Ha et al. (2012) designed a weather monitoring system based off current spatial and temporal methods which facilitates spatiotemporal queries and stores sensor data. The system consisted of two main methods, Time-Segment Insertion (TSI) and Time-Point Insertion (TPI), which save storage space without losing raw data as only one sample record (i.e. a tuple) needs to be stored in memory when it copies the values acquired over some-interval, be that time or otherwise. Ha et al. (2012) also reduced costs of processing spatiotemporal queries through this WSN by filtering out unneeded sensors from a range with support from a fixed grid which identified the sensors location. Innovations which reduce cost and save space would be essential for scaling up and scaling out additional work on this research project.