*3.4. Heterocyst Frequency*

To determine if this effect on N2-fixation was perhaps due to an inability to develop heterocysts (a developmentally specialized cell type dedicated to this process), we conducted microscopic counts of vegetative cells and heterocysts in strains incubated for seven days at different temperatures (Table 1). Counts were performed only on apparently healthy filaments, but at 35 and 40 ◦C, biomass from all replicates of *Nostoc* sp. HSN008 and *Tolypothrix* sp. HSN042 looked yellowish, and microscopy revealed high cell mortality as well. In fact, in one occasion, one set of replicates of *Nostoc* sp. HSN008 did not survive to day 7 (Table 1). All strains grown at 25 ◦C looked healthy when counts were performed. Those caveats aside, the frequency of heterocysts declined precipitously for *Nostoc* spp. strains above 35 degrees, and above 30 degrees for *Tolypothrix* spp. strains. In *Scytonema* spp., there were only slight decreases in this frequency in the temperature range tested. This is consistent with a cell developmental basis for the sensitivity of N2-fixation to high temperatures in *Nostoc* spp. and *Tolypothrix* spp.


**Table 1.** Frequency of heterocysts (number of vegetative cells per heterocyst) in representative cyanobacterial strains after incubation at 30, 35 and 40 ◦C for 7 days. Averages of *n* = 6 determinations ± standard deviation are given. H: heterocystous, VG: Vegetative cells.

#### *3.5. Thermal Niche of Biocrust Heterocystous Cyanobacteria through Meta-Analyses of Molecular Surveys*

A total of 84 locations from eleven di fferent biocrust surveys conducted in di fferent arid and semiarid regions in North and South America, Europe, Australia, and China, and in the Brazil Savannah (see Table S4), were used in a meta-analysis to assess the relative contribution of the three main clades of heterocystous cyanobacteria along temperature related parameters. Figure 4 shows the relative proportion of *Scytonema* spp., *Nostoc* spp. and *Tolypothrix* spp., plotted against the mean annual temperature (MAT) of origin and the corresponding mean temperature during the wettest quarter of the year (MTempWetQ). MTempWetQ was used as a proxy for growth season since biocrust organisms are metabolically active only when water is available [65] and are relatively insensitive to heat stress when dry. The relative abundance of *Scytonema* spp. was positively correlated with MAT (*p* = 4 × <sup>10</sup>−4) and with MTempWetQ (*p* = <sup>10</sup>−8); *Nostoc* spp. was negatively correlated with MAT (*p* = 5 × <sup>10</sup>−3) and with MTempWetQ (*p*= <sup>10</sup>−7). The relative abundance of *Tolypothrix* spp. was also negatively correlated with MAT (*p*= 0.035) and with MTempWetQ (*p* = 3 × <sup>10</sup>−3). Using linear regression on arcsine transformed data, MAT explained 15, 9 and 1 of the variability in *Scytonema* spp., *Nostoc* spp. and *Tolypothrix* spp., respectively. The explanatory power of MTempWetQ was much higher in all cases, rising to 32, 28 and 11%, respectively. Based on MTempWetQ, *Scytonema* spp. could attain dominance at warmer temperatures (Figure 4A), while at lower temperatures, *Tolypothrix* spp. (Figure 4C), followed by *Nostoc* spp. (Figure 4B) attain higher maximal relative abundances. Detailed statistics are in Supplementary Materials (Figure S1, Tables S5–S10).

**Figure 4.** Proportion of sequence reads assignable to *Scytonema* spp. (**A**; orange), *Nostoc* spp. (**B**; yellow), and *Tolypothrix* spp. (**C**; blue) to those assignable to all heterocystous (Order Nostocales) cyanobacteria, in 16S rRNA molecular survey datasets, as a function of climate temperature indicators. Data are from biocrust communities surveyed at 84 locations around the world (see Table S4). Each dot represents a different location.
