Temperature, salinity and pH of the water column and pore water during a nutrient enrichment and macrofauna exclusion experiment. A total of 24 plots were set up parallel to the shore, with presence of the three seagrass species (Syringodium isoetifolium, Thalassodendron ciliatum and Thalassia hemprichii) in each plot. The experiment was the factorial combination of two treatments: macrofauna exclusion using cages (three levels: open, closed and uncaged) and nutrient enrichment using garden NPK fertilizer (two levels: ambient and enriched). Each treatment combination was replicated four times. The macrofauna exclusion cages were made of wooden pegs (0.5 x 0.5m wide and 0.5 m high) and were covered with a plastic net (14 mm diagonal mesh size). This mesh-size excludes macrofauna, but not small fish and mesofauna. This macrofauna exclusion technique (Casini et al., 2008) simulates the consequences for the food web of losing a top predators and macrograzers. Open cages had all the side meshes halved to allow partial access from all four sides (to check for cage effects), and uncaged plots were marked with four wooden pegs. The cages were anchored by metal T bars and cleaned every 2 weeks to remove organisms and filamentous algae that would otherwise affect light penetration and water flow. Nutrient enrichment was simulated by issuing nitrogen-phosphorus-potassium (NPK) [15:9:20] fertilizer pellets. NPK fertilizer fast release pellets were packed into cotton tubes and then into a perforated plastic tubes to simulate slow release of nutrients. The tubes were filled with 75g of pellets and five tubes were placed in each of the enriched plots (375g in total per plot) following the protocol by Heck et al. (2000). The tubes were buried half-way into the sediment to ensure enrichment of both the water column and the sediment. Fertilizer was replaced every 3 weeks, making a total of 1125 grams of fertilizer released per enriched plot at the end of the experiment. All plots were separated by at least 3 metres, proven as a no-overspill area in previous nutrient enrichment experiments (Lotze & Worm, 2000). Data was collected between July 19th and September 20th of 2017 in four sampling times. Day 0 (19.07.2017), Day 20 (09.08.2017), Day 38 (31.08.2017) and Day 63 (19.09.2017). Data collection and experiment took place in Changuu Island (Zanzibar Archipelago, Tanzania; 06˚11'S, 39˚16'E). Changu Island is located 3 km from Stone Town, Zanzibar's busiest town. Changuu remains relatively unaffected by nutrient runoff pollution, although it has a small tourist resort and a large tortoise centre with daily boat traffic bringing tourists from Stone town. The study area is characterised by a fringing reef around a multi-specific seagrass ecosystem. The substrate is primarily carbonate sediment. Average water depth is approximately 30 cm at Spring Low and 5 m at Spring high tide with an average depth of 2 m. General environmental parameters were collected at each sampling time for accounting for potential confounding variables during the experiment. A Multiprobe 3430 (WTW GmbH) was utilized to measure temperature (TetraCon 925, ̊C), salinity (TetraCon 925, psu) and pH (SenTix 950). Measurements were taken in the morning approximately every two hours at each sampling time.